27 april 2026

Een open-source spec voor Codex-orkestratie: Symphony

Door Alex Kotliarskyi, Victor Zhu en Zach Brock

Bezig met laden...

Zes maanden geleden nam ons team, terwijl we aan een interne productiviteitstool werkten, een beslissing die toen controversieel was: we zouden onze repo bouwen zonder door mensen geschreven code. Elke regel in onze projectrepository moest door Codex worden gegenereerd.

Om dat te laten werken, hebben we onze engineeringworkflow vanaf de basis opnieuw ontworpen. We bouwden een agent-vriendelijke repository, investeerden zwaar in geautomatiseerde tests en guardrails, en behandelden Codex als een volwaardige teamgenoot. We documenteerden die reis in onze eerdere blogpost over harness engineering⁠.

Het werkte, maar we liepen vervolgens tegen een nieuwe bottleneck aan: contextwisseling.

Om dit nieuwe probleem op te lossen, bouwden we een systeem genaamd Symphony. Symphony⁠(opent in een nieuw venster) is een agent-orkestrator die een projectmanagementtool zoals Linear verandert in een control plane voor programmeeragents. Elke open taak krijgt een agent, agents draaien continu en mensen beoordelen de resultaten.

In deze post leggen we uit hoe we Symphony hebben gemaakt, wat bij sommige teams leidde tot een stijging van 500% in gemergede pull requests, en hoe je het kunt gebruiken om van je eigen issuetracker een altijd actieve agent-orkestrator te maken.

De grenzen van interactieve programmeeragents

Zelfs nu programmeeragents steeds makkelijker te gebruiken zijn, blijven het interactieve tools, of je ze nu via webapps of de CLI gebruikt.

Naarmate er bij OpenAI meer agent-ondersteund werk kwam, liepen we tegen een nieuwe uitdaging aan. Elke engineer opende een paar Codex-sessies, wees taken toe, beoordeelde de output, stuurde de agent bij en herhaalde dat proces. In de praktijk konden de meeste mensen zonder al te veel moeite drie tot vijf sessies tegelijk aan voordat het voortdurende schakelen tussen contexten te veel werd. Daarboven nam de productiviteit af. We vergaten welke sessie waarmee bezig was, sprongen tussen terminals om agents weer op koers te brengen en moesten problemen oplossen bij langlopende taken die halverwege vastliepen.

De agents waren snel, maar we hadden een systeembottleneck: menselijke aandacht. We hadden in feite een team van extreem capabele junior engineers opgebouwd en vervolgens onze menselijke engineers de taak gegeven hen te micromanagen. Dat is onmogelijk op te schalen.

Een verschuiving in perspectief

We beseften dat we op het verkeerde optimaliseerden. We richtten ons systeem op programmeersessies en gemergede PR's, terwijl PR's en sessies eigenlijk slechts tussenstappen zijn. Softwareworkflows zijn grotendeels ingericht rond concrete opleveringen: issues, taken, tickets en mijlpalen.

Dus vroegen we ons af wat er zou gebeuren als we agents niet langer direct zouden aansturen en ze in plaats daarvan werk uit onze taaktracker zouden laten halen.

Dat idee werd Symphony, een geschreven spec die als supervisor functioneert om agentic werk te coördineren.

Hoe we van onze issuetracker een agent-orkestrator maakten

Symphony begon met een eenvoudig concept: elke open taak moet door een agent worden opgepakt en afgerond. In plaats van Codex-sessies in meerdere tabbladen te beheren, maakten we van onze issuetracker de control plane.

In deze opzet wordt elk open Linear-issue gekoppeld aan een specifieke agent-werkruimte. Symphony houdt het takenbord voortdurend in de gaten en zorgt ervoor dat voor elke actieve taak een agent blijft draaien totdat die is afgerond. Als een agent crasht of vastloopt, start Symphony die opnieuw. Als er nieuw werk verschijnt, pakt Symphony dat op en begint het te organiseren.

We hebben onze workflow opgebouwd rond ticketstatussen, waarbij we taakbeheerder Linear als een statusmachine gebruikten.

Coding agents use Linear as a state machine to work alongside us.

In de praktijk ontkoppelt Symphony werk van sessies en van pull requests. Sommige issues leveren meerdere PR's op in verschillende repo's; andere zijn puur onderzoek of analyse en raken de codebase nooit.

Zodra werk op deze manier abstract wordt gemaakt, kunnen tickets veel grotere eenheden werk vertegenwoordigen.

We gebruiken Symphony regelmatig om complexe features en infrastructuurmigraties te coördineren. We kunnen bijvoorbeeld een taak indienen waarin we de agent vragen de codebase, Slack of Notion te analyseren en een implementatieplan op te stellen. Zodra we tevreden zijn met het plan, genereert de agent een taakstructuur waarin het werk in fasen wordt opgesplitst en afhankelijkheden tussen taken worden vastgelegd.

Agents beginnen alleen aan taken die niet geblokkeerd zijn, dus de uitvoering ontvouwt zich voor deze DAG (een reeks uitvoeringsstappen) op natuurlijke en optimale wijze parallel. In het onderstaande voorbeeld markeerden we de React-upgrade als geblokkeerd door een migratie naar Vite. Zoals verwacht begonnen agents pas met het upgraden van React nadat de migratie naar Vite was voltooid.

Agents kunnen ook zelf werk creëren. Tijdens implementatie of review merken ze vaak verbeteringen op die buiten de scope van de huidige taak vallen: een performanceprobleem, een kans op refactoring of een betere architectuur. Als dat gebeurt, dienen ze simpelweg een nieuw issue in dat we later kunnen beoordelen en inplannen. Veel van deze vervolgtaken worden ook door agents opgepakt. Terwijl wij toezicht houden op dit proces, blijven agents georganiseerd en houden ze het werk in beweging.

Deze manier van werken verlaagt de mentale drempel om aan onduidelijk werk te beginnen aanzienlijk. Als de agent iets verkeerd doet, levert dat nog steeds nuttige informatie op, terwijl het ons bijna niets kost. We kunnen heel eenvoudig tickets indienen zodat de agent prototypes maakt en verschillende richtingen verkent, en alles wat niet bevalt weer verwerpen.

Omdat de orkestrator op devboxes draait en nooit slaapt, kunnen we overal taken toevoegen en weten dat een agent ze oppakt. Zo voerde een engineer in ons team drie belangrijke wijzigingen door vanuit de Linear-app op zijn telefoon, vanuit een vakantiehuisje met slechte wifi.

Meer verkenning door op deze manier te werken

Toen we de effecten van werken met Symphony zagen, was de meest opvallende verandering de output. Bij sommige teams binnen OpenAI zagen we het aantal gemergede PR's in de eerste drie weken met een factor zes toenemen. Buiten OpenAI wees Linear-oprichter Karri Saarinen op een piek in aangemaakte werkruimtes⁠(opent in een nieuw venster) toen we Symphony uitbrachten. De grotere verschuiving zit echter in hoe teams over werk denken.

Wanneer onze engineers niet langer tijd besteden aan het begeleiden van Codex-sessies, veranderen de economische verhoudingen van codewijzigingen volledig. De kosten van elke wijziging dalen, omdat we niet langer menselijke inspanning steken in het aansturen van de implementatie zelf.

Dat veranderde ons gedrag. Het is heel eenvoudig geworden om speculatieve taken in Symphony op te starten. Probeer een idee uit, verken een refactor, test een hypothese en ga alleen verder met resultaten die veelbelovend lijken.

Het vergroot ook de groep mensen die werk kan initiëren. Onze productmanager en designer kunnen nu featureverzoeken rechtstreeks in Symphony indienen. Ze hoeven niet de repo uit te checken of een Codex-sessie te beheren. Ze beschrijven de feature en krijgen een reviewpakket terug dat een videowalkthrough bevat van de feature die in het echte product werkt.

Symphony blinkt ook uit in grote monorepo's (zoals die we bij OpenAI hebben), waar de laatste stap om een PR te mergen traag en kwetsbaar kan zijn. Het systeem bewaakt CI, rebaset wanneer nodig, lost conflicten op, probeert flaky checks opnieuw en loodst wijzigingen door de pijplijn. Tegen de tijd dat een ticket Merging bereikt, hebben we veel vertrouwen dat de wijziging zonder handmatig ingrijpen in de hoofdbranch terechtkomt.

Vooruitgang brengt nieuwe problemen met zich mee

Werken op dit niveau brengt trade-offs met zich mee. Toen we overstapten van het interactief aansturen van agents naar het toewijzen van werk op ticketniveau, verloren we de mogelijkheid om ze onderweg voortdurend bij te sturen en waar nodig van koers te laten veranderen. Soms leverde de agent iets op dat de plank volledig missloeg. Dat was nuttig: zulke missers brachten zwakke plekken in het systeem aan het licht en hielpen ons het robuuster te maken.

In plaats van het resultaat handmatig te patchen, voegden we guardrails en skills toe zodat de agents de volgende keer beter werk konden leveren. Na verloop van tijd leidde dit ertoe dat we nieuwe mogelijkheden aan onze harness toevoegden, zoals end-to-endtests uitvoeren, de app aansturen via Chrome DevTools en QA-smoketests beheren. We verbeterden onze documentatie aanzienlijk en maakten duidelijker wat we onder goed werk verstaan.

Niet elke taak past bij hoe Symphony werkt. Sommige problemen vereisen nog steeds softwareontwikkelaars die rechtstreeks werken met interactieve Codex-sessies, vooral ambigue problemen of werk dat sterk oordeel en expertise vereist. In de praktijk zijn dit meestal de interessantste en prettigste taken voor onze engineers om tijd aan te besteden.

Het verschil is dat Symphony het grootste deel van routinematig implementatiewerk aankan. Daardoor kunnen engineers zich op één moeilijk probleem tegelijk richten in plaats van voortdurend te schakelen tussen kleinere taken.

We leerden ook dat het niet goed werkt om agents te behandelen als rigide stappen in een statusmachine. Modellen worden slimmer en kunnen grotere problemen oplossen dan de kaders die wij ze opleggen. In de vroege versies maakten alle GitHub-integraties bijvoorbeeld deel uit van de harness eromheen: Codex werd alleen ingezet om codewijzigingen door te voeren, terwijl de rest van het proces, zoals wijzigingen indienen en tests uitvoeren, in code was vastgelegd. In die vroege vorm van agent-ondersteund werk vroegen we Codex dus alleen om de taak te implementeren. Die aanpak bleek te beperkend. Codex kan prima meerdere PR's aanmaken, reviewfeedback lezen en verwerken. Daarom gaven we het meer tools: de gh-CLI, skills om CI-logs te lezen, enzovoort. Nu kunnen we Codex ook vragen om meer te doen, zoals oude PR's sluiten of rapporten ophalen over afgerond versus stopgezet werk. Zulke taken vallen ver buiten de oorspronkelijke focus op feature-implementatie.

Daarom zijn we uiteindelijk overgestapt op het geven van doelstellingen aan agents in plaats van strikte overgangen, net zoals een goede manager een doel toewijst aan iemand in zijn team. De kracht van modellen komt voort uit hun vermogen tot redenering, dus geef ze tools en context en laat ze hun werk doen.

Symphony gebruiken om Symphony te bouwen

Wanneer je de Symphony-repository opent, is het eerste wat je opvalt dat Symphony technisch gezien gewoon een SPEC.md-bestand is: een definitie van het probleem en de beoogde oplossing. In plaats van een complex supervisiesysteem te bouwen, definieerden we het probleem en de beoogde oplossingen en gaven we agents sturing op hoog niveau.

Markdown

1# Symphony Service Specification
2
3Status: Draft v1 (language-agnostic)
4
5Purpose: Define a service that orchestrates coding agents to get project work done.
6
7## 1. Problem Statement
8
9Symphony is a long-running automation service that continuously reads work from an issue tracker
10(Linear in this specification version), creates an isolated workspace for each issue, and runs a
11coding agent session for that issue inside the workspace.
12
13The service solves four operational problems:
14
15- It turns issue execution into a repeatable daemon workflow instead of manual scripts.
16- It isolates agent execution in per-issue workspaces so agent commands run only inside per-issue
17  workspace directories.
18- It keeps the workflow policy in-repo (`WORKFLOW.md`) so teams version the agent prompt and runtime
19  settings with their code.
20- It provides enough observability to operate and debug multiple concurrent agent runs.
21
22Implementations are expected to document their trust and safety posture explicitly. This
23specification does not require a single approval, sandbox, or operator-confirmation policy; some
24implementations may target trusted environments with a high-trust configuration, while others may
25require stricter approvals or sandboxing.
26
27Important boundary:
28
29- Symphony is a scheduler/runner and tracker reader.
30- Ticket writes (state transitions, comments, PR links) are typically performed by the coding agent
31  using tools available in the workflow/runtime environment.
32- A successful run may end at a workflow-defined handoff state (for example `Human Review`), not
33  necessarily `Done`.
34
35## 2. Goals and Non-Goals
36
37### 2.1 Goals
38
39- Poll the issue tracker on a fixed cadence and dispatch work with bounded concurrency.
40- Maintain a single authoritative orchestrator state for dispatch, retries, and reconciliation.
41- Create deterministic per-issue workspaces and preserve them across runs.
42- Stop active runs when issue state changes make them ineligible.
43- Recover from transient failures with exponential backoff.
44- Load runtime behavior from a repository-owned `WORKFLOW.md` contract.
45- Expose operator-visible observability (at minimum structured logs).
46- Support restart recovery without requiring a persistent database.
47
48### 2.2 Non-Goals
49
50- Rich web UI or multi-tenant control plane.
51- Prescribing a specific dashboard or terminal UI implementation.
52- General-purpose workflow engine or distributed job scheduler.
53- Built-in business logic for how to edit tickets, PRs, or comments. (That logic lives in the
54  workflow prompt and agent tooling.)
55- Mandating strong sandbox controls beyond what the coding agent and host OS provide.
56- Mandating a single default approval, sandbox, or operator-confirmation posture for all
57  implementations.
58
59## 3. System Overview
60
61### 3.1 Main Components
62
631. `Workflow Loader`
64   - Reads `WORKFLOW.md`.
65   - Parses YAML front matter and prompt body.
66   - Returns `{config, prompt_template}`.
67
682. `Config Layer`
69   - Exposes typed getters for workflow config values.
70   - Applies defaults and environment variable indirection.
71   - Performs validation used by the orchestrator before dispatch.
72
733. `Issue Tracker Client`
74   - Fetches candidate issues in active states.
75   - Fetches current states for specific issue IDs (reconciliation).
76   - Fetches terminal-state issues during startup cleanup.
77   - Normalizes tracker payloads into a stable issue model.
78
794. `Orchestrator`
80   - Owns the poll tick.
81   - Owns the in-memory runtime state.
82   - Decides which issues to dispatch, retry, stop, or release.
83   - Tracks session metrics and retry queue state.
84
855. `Workspace Manager`
86   - Maps issue identifiers to workspace paths.
87   - Ensures per-issue workspace directories exist.
88   - Runs workspace lifecycle hooks.
89   - Cleans workspaces for terminal issues.
90
916. `Agent Runner`
92   - Creates workspace.
93   - Builds prompt from issue + workflow template.
94   - Launches the coding agent app-server client.
95   - Streams agent updates back to the orchestrator.
96
977. `Status Surface` (optional)
98   - Presents human-readable runtime status (for example terminal output, dashboard, or other
99     operator-facing view).
100
1018. `Logging`
102   - Emits structured runtime logs to one or more configured sinks.
103
104### 3.2 Abstraction Levels
105
106Symphony is easiest to port when kept in these layers:
107
1081. `Policy Layer` (repo-defined)
109   - `WORKFLOW.md` prompt body.
110   - Team-specific rules for ticket handling, validation, and handoff.
111
1122. `Configuration Layer` (typed getters)
113   - Parses front matter into typed runtime settings.
114   - Handles defaults, environment tokens, and path normalization.
115
1163. `Coordination Layer` (orchestrator)
117   - Polling loop, issue eligibility, concurrency, retries, reconciliation.
118
1194. `Execution Layer` (workspace + agent subprocess)
120   - Filesystem lifecycle, workspace preparation, coding-agent protocol.
121
1225. `Integration Layer` (Linear adapter)
123   - API calls and normalization for tracker data.
124
1256. `Observability Layer` (logs + optional status surface)
126   - Operator visibility into orchestrator and agent behavior.
127
128### 3.3 External Dependencies
129
130- Issue tracker API (Linear for `tracker.kind: linear` in this specification version).
131- Local filesystem for workspaces and logs.
132- Optional workspace population tooling (for example Git CLI, if used).
133- Coding-agent executable that supports JSON-RPC-like app-server mode over stdio.
134- Host environment authentication for the issue tracker and coding agent.
135
136## 4. Core Domain Model
137
138### 4.1 Entities
139
140#### 4.1.1 Issue
141
142Normalized issue record used by orchestration, prompt rendering, and observability output.
143
144Fields:
145
146- `id` (string)
147  - Stable tracker-internal ID.
148- `identifier` (string)
149  - Human-readable ticket key (example: `ABC-123`).
150- `title` (string)
151- `description` (string or null)
152- `priority` (integer or null)
153  - Lower numbers are higher priority in dispatch sorting.
154- `state` (string)
155  - Current tracker state name.
156- `branch_name` (string or null)
157  - Tracker-provided branch metadata if available.
158- `url` (string or null)
159- `labels` (list of strings)
160  - Normalized to lowercase.
161- `blocked_by` (list of blocker refs)
162  - Each blocker ref contains:
163    - `id` (string or null)
164    - `identifier` (string or null)
165    - `state` (string or null)
166- `created_at` (timestamp or null)
167- `updated_at` (timestamp or null)
168
169#### 4.1.2 Workflow Definition
170
171Parsed `WORKFLOW.md` payload:
172
173- `config` (map)
174  - YAML front matter root object.
175- `prompt_template` (string)
176  - Markdown body after front matter, trimmed.
177
178#### 4.1.3 Service Config (Typed View)
179
180Typed runtime values derived from `WorkflowDefinition.config` plus environment resolution.
181
182Examples:
183
184- poll interval
185- workspace root
186- active and terminal issue states
187- concurrency limits
188- coding-agent executable/args/timeouts
189- workspace hooks
190
191#### 4.1.4 Workspace
192
193Filesystem workspace assigned to one issue identifier.
194
195Fields (logical):
196
197- `path` (workspace path; current runtime typically uses absolute paths, but relative roots are
198  possible if configured without path separators)
199- `workspace_key` (sanitized issue identifier)
200- `created_now` (boolean, used to gate `after_create` hook)
201
202#### 4.1.5 Run Attempt
203
204One execution attempt for one issue.
205
206Fields (logical):
207
208- `issue_id`
209- `issue_identifier`
210- `attempt` (integer or null, `null` for first run, `>=1` for retries/continuation)
211- `workspace_path`
212- `started_at`
213- `status`
214- `error` (optional)
215
216#### 4.1.6 Live Session (Agent Session Metadata)
217
218State tracked while a coding-agent subprocess is running.
219
220Fields:
221
222- `session_id` (string, `<thread_id>-<turn_id>`)
223- `thread_id` (string)
224- `turn_id` (string)
225- `codex_app_server_pid` (string or null)
226- `last_codex_event` (string/enum or null)
227- `last_codex_timestamp` (timestamp or null)
228- `last_codex_message` (summarized payload)
229- `codex_input_tokens` (integer)
230- `codex_output_tokens` (integer)
231- `codex_total_tokens` (integer)
232- `last_reported_input_tokens` (integer)
233- `last_reported_output_tokens` (integer)
234- `last_reported_total_tokens` (integer)
235- `turn_count` (integer)
236  - Number of coding-agent turns started within the current worker lifetime.
237
238#### 4.1.7 Retry Entry
239
240Scheduled retry state for an issue.
241
242Fields:
243
244- `issue_id`
245- `identifier` (best-effort human ID for status surfaces/logs)
246- `attempt` (integer, 1-based for retry queue)
247- `due_at_ms` (monotonic clock timestamp)
248- `timer_handle` (runtime-specific timer reference)
249- `error` (string or null)
250
251#### 4.1.8 Orchestrator Runtime State
252
253Single authoritative in-memory state owned by the orchestrator.
254
255Fields:
256
257- `poll_interval_ms` (current effective poll interval)
258- `max_concurrent_agents` (current effective global concurrency limit)
259- `running` (map `issue_id -> running entry`)
260- `claimed` (set of issue IDs reserved/running/retrying)
261- `retry_attempts` (map `issue_id -> RetryEntry`)
262- `completed` (set of issue IDs; bookkeeping only, not dispatch gating)
263- `codex_totals` (aggregate tokens + runtime seconds)
264- `codex_rate_limits` (latest rate-limit snapshot from agent events)
265
266### 4.2 Stable Identifiers and Normalization Rules
267
268- `Issue ID`
269  - Use for tracker lookups and internal map keys.
270- `Issue Identifier`
271  - Use for human-readable logs and workspace naming.
272- `Workspace Key`
273  - Derive from `issue.identifier` by replacing any character not in `[A-Za-z0-9._-]` with `_`.
274  - Use the sanitized value for the workspace directory name.
275- `Normalized Issue State`
276  - Compare states after `lowercase`.
277- `Session ID`
278  - Compose from coding-agent `thread_id` and `turn_id` as `<thread_id>-<turn_id>`.
279
280## 5. Workflow Specification (Repository Contract)
281
282### 5.1 File Discovery and Path Resolution
283
284Workflow file path precedence:
285
2861. Explicit application/runtime setting (set by CLI startup path).
2872. Default: `WORKFLOW.md` in the current process working directory.
288
289Loader behavior:
290
291- If the file cannot be read, return `missing_workflow_file` error.
292- The workflow file is expected to be repository-owned and version-controlled.
293
294### 5.2 File Format
295
296`WORKFLOW.md` is a Markdown file with optional YAML front matter.
297
298Design note:
299
300- `WORKFLOW.md` should be self-contained enough to describe and run different workflows (prompt,
301  runtime settings, hooks, and tracker selection/config) without requiring out-of-band
302  service-specific configuration.
303
304Parsing rules:
305
306- If file starts with `---`, parse lines until the next `---` as YAML front matter.
307- Remaining lines become the prompt body.
308- If front matter is absent, treat the entire file as prompt body and use an empty config map.
309- YAML front matter must decode to a map/object; non-map YAML is an error.
310- Prompt body is trimmed before use.
311
312Returned workflow object:
313
314- `config`: front matter root object (not nested under a `config` key).
315- `prompt_template`: trimmed Markdown body.
316
317### 5.3 Front Matter Schema
318
319Top-level keys:
320
321- `tracker`
322- `polling`
323- `workspace`
324- `hooks`
325- `agent`
326- `codex`
327
328Unknown keys should be ignored for forward compatibility.
329
330Note:
331
332- The workflow front matter is extensible. Optional extensions may define additional top-level keys
333  (for example `server`) without changing the core schema above.
334- Extensions should document their field schema, defaults, validation rules, and whether changes
335  apply dynamically or require restart.
336- Common extension: `server.port` (integer) enables the optional HTTP server described in Section
337  13.7.
338
339#### 5.3.1 `tracker` (object)
340
341Fields:
342
343- `kind` (string)
344  - Required for dispatch.
345  - Current supported value: `linear`
346- `endpoint` (string)
347  - Default for `tracker.kind == "linear"`: `https://api.linear.app/graphql`
348- `api_key` (string)
349  - May be a literal token or `$VAR_NAME`.
350  - Canonical environment variable for `tracker.kind == "linear"`: `LINEAR_API_KEY`.
351  - If `$VAR_NAME` resolves to an empty string, treat the key as missing.
352- `project_slug` (string)
353  - Required for dispatch when `tracker.kind == "linear"`.
354- `active_states` (list of strings)
355  - Default: `Todo`, `In Progress`
356- `terminal_states` (list of strings)
357  - Default: `Closed`, `Cancelled`, `Canceled`, `Duplicate`, `Done`
358
359#### 5.3.2 `polling` (object)
360
361Fields:
362
363- `interval_ms` (integer or string integer)
364  - Default: `30000`
365  - Changes should be re-applied at runtime and affect future tick scheduling without restart.
366
367#### 5.3.3 `workspace` (object)
368
369Fields:
370
371- `root` (path string or `$VAR`)
372  - Default: `<system-temp>/symphony_workspaces`
373  - `~` and strings containing path separators are expanded.
374  - Bare strings without path separators are preserved as-is (relative roots are allowed but
375    discouraged).
376
377#### 5.3.4 `hooks` (object)
378
379Fields:
380
381- `after_create` (multiline shell script string, optional)
382  - Runs only when a workspace directory is newly created.
383  - Failure aborts workspace creation.
384- `before_run` (multiline shell script string, optional)
385  - Runs before each agent attempt after workspace preparation and before launching the coding
386    agent.
387  - Failure aborts the current attempt.
388- `after_run` (multiline shell script string, optional)
389  - Runs after each agent attempt (success, failure, timeout, or cancellation) once the workspace
390    exists.
391  - Failure is logged but ignored.
392- `before_remove` (multiline shell script string, optional)
393  - Runs before workspace deletion if the directory exists.
394  - Failure is logged but ignored; cleanup still proceeds.
395- `timeout_ms` (integer, optional)
396  - Default: `60000`
397  - Applies to all workspace hooks.
398  - Non-positive values should be treated as invalid and fall back to the default.
399  - Changes should be re-applied at runtime for future hook executions.
400
401#### 5.3.5 `agent` (object)
402
403Fields:
404
405- `max_concurrent_agents` (integer or string integer)
406  - Default: `10`
407  - Changes should be re-applied at runtime and affect subsequent dispatch decisions.
408- `max_retry_backoff_ms` (integer or string integer)
409  - Default: `300000` (5 minutes)
410  - Changes should be re-applied at runtime and affect future retry scheduling.
411- `max_concurrent_agents_by_state` (map `state_name -> positive integer`)
412  - Default: empty map.
413  - State keys are normalized (`lowercase`) for lookup.
414  - Invalid entries (non-positive or non-numeric) are ignored.
415
416#### 5.3.6 `codex` (object)
417
418Fields:
419
420For Codex-owned config values such as `approval_policy`, `thread_sandbox`, and
421`turn_sandbox_policy`, supported values are defined by the targeted Codex app-server version.
422Implementors should treat them as pass-through Codex config values rather than relying on a
423hand-maintained enum in this spec. To inspect the installed Codex schema, run
424`codex app-server generate-json-schema --out <dir>` and inspect the relevant definitions referenced
425by `v2/ThreadStartParams.json` and `v2/TurnStartParams.json`. Implementations may validate these
426fields locally if they want stricter startup checks.
427
428- `command` (string shell command)
429  - Default: `codex app-server`
430  - The runtime launches this command via `bash -lc` in the workspace directory.
431  - The launched process must speak a compatible app-server protocol over stdio.
432- `approval_policy` (Codex `AskForApproval` value)
433  - Default: implementation-defined.
434- `thread_sandbox` (Codex `SandboxMode` value)
435  - Default: implementation-defined.
436- `turn_sandbox_policy` (Codex `SandboxPolicy` value)
437  - Default: implementation-defined.
438- `turn_timeout_ms` (integer)
439  - Default: `3600000` (1 hour)
440- `read_timeout_ms` (integer)
441  - Default: `5000`
442- `stall_timeout_ms` (integer)
443  - Default: `300000` (5 minutes)
444  - If `<= 0`, stall detection is disabled.
445
446### 5.4 Prompt Template Contract
447
448The Markdown body of `WORKFLOW.md` is the per-issue prompt template.
449
450Rendering requirements:
451
452- Use a strict template engine (Liquid-compatible semantics are sufficient).
453- Unknown variables must fail rendering.
454- Unknown filters must fail rendering.
455
456Template input variables:
457
458- `issue` (object)
459  - Includes all normalized issue fields, including labels and blockers.
460- `attempt` (integer or null)
461  - `null`/absent on first attempt.
462  - Integer on retry or continuation run.
463
464Fallback prompt behavior:
465
466- If the workflow prompt body is empty, the runtime may use a minimal default prompt
467  (`You are working on an issue from Linear.`).
468- Workflow file read/parse failures are configuration/validation errors and should not silently fall
469  back to a prompt.
470
471### 5.5 Workflow Validation and Error Surface
472
473Error classes:
474
475- `missing_workflow_file`
476- `workflow_parse_error`
477- `workflow_front_matter_not_a_map`
478- `template_parse_error` (during prompt rendering)
479- `template_render_error` (unknown variable/filter, invalid interpolation)
480
481Dispatch gating behavior:
482
483- Workflow file read/YAML errors block new dispatches until fixed.
484- Template errors fail only the affected run attempt.
485
486## 6. Configuration Specification
487
488### 6.1 Source Precedence and Resolution Semantics
489
490Configuration precedence:
491
4921. Workflow file path selection (runtime setting -> cwd default).
4932. YAML front matter values.
4943. Environment indirection via `$VAR_NAME` inside selected YAML values.
4954. Built-in defaults.
496
497Value coercion semantics:
498
499- Path/command fields support:
500  - `~` home expansion
501  - `$VAR` expansion for env-backed path values
502  - Apply expansion only to values intended to be local filesystem paths; do not rewrite URIs or
503    arbitrary shell command strings.
504
505### 6.2 Dynamic Reload Semantics
506
507Dynamic reload is required:
508
509- The software should watch `WORKFLOW.md` for changes.
510- On change, it should re-read and re-apply workflow config and prompt template without restart.
511- The software should attempt to adjust live behavior to the new config (for example polling
512  cadence, concurrency limits, active/terminal states, codex settings, workspace paths/hooks, and
513  prompt content for future runs).
514- Reloaded config applies to future dispatch, retry scheduling, reconciliation decisions, hook
515  execution, and agent launches.
516- Implementations are not required to restart in-flight agent sessions automatically when config
517  changes.
518- Extensions that manage their own listeners/resources (for example an HTTP server port change) may
519  require restart unless the implementation explicitly supports live rebind.
520- Implementations should also re-validate/reload defensively during runtime operations (for example
521  before dispatch) in case filesystem watch events are missed.
522- Invalid reloads should not crash the service; keep operating with the last known good effective
523  configuration and emit an operator-visible error.
524
525### 6.3 Dispatch Preflight Validation
526
527This validation is a scheduler preflight run before attempting to dispatch new work. It validates
528the workflow/config needed to poll and launch workers, not a full audit of all possible workflow
529behavior.
530
531Startup validation:
532
533- Validate configuration before starting the scheduling loop.
534- If startup validation fails, fail startup and emit an operator-visible error.
535
536Per-tick dispatch validation:
537
538- Re-validate before each dispatch cycle.
539- If validation fails, skip dispatch for that tick, keep reconciliation active, and emit an
540  operator-visible error.
541
542Validation checks:
543
544- Workflow file can be loaded and parsed.
545- `tracker.kind` is present and supported.
546- `tracker.api_key` is present after `$` resolution.
547- `tracker.project_slug` is present when required by the selected tracker kind.
548- `codex.command` is present and non-empty.
549
550### 6.4 Config Fields Summary (Cheat Sheet)
551
552This section is intentionally redundant so a coding agent can implement the config layer quickly.
553
554- `tracker.kind`: string, required, currently `linear`
555- `tracker.endpoint`: string, default `https://api.linear.app/graphql` when `tracker.kind=linear`
556- `tracker.api_key`: string or `$VAR`, canonical env `LINEAR_API_KEY` when `tracker.kind=linear`
557- `tracker.project_slug`: string, required when `tracker.kind=linear`
558- `tracker.active_states`: list of strings, default `["Todo", "In Progress"]`
559- `tracker.terminal_states`: list of strings, default `["Closed", "Cancelled", "Canceled", "Duplicate", "Done"]`
560- `polling.interval_ms`: integer, default `30000`
561- `workspace.root`: path, default `<system-temp>/symphony_workspaces`
562- `worker.ssh_hosts` (extension): list of SSH host strings, optional; when omitted, work runs
563  locally
564- `worker.max_concurrent_agents_per_host` (extension): positive integer, optional; shared per-host
565  cap applied across configured SSH hosts
566- `hooks.after_create`: shell script or null
567- `hooks.before_run`: shell script or null
568- `hooks.after_run`: shell script or null
569- `hooks.before_remove`: shell script or null
570- `hooks.timeout_ms`: integer, default `60000`
571- `agent.max_concurrent_agents`: integer, default `10`
572- `agent.max_turns`: integer, default `20`
573- `agent.max_retry_backoff_ms`: integer, default `300000` (5m)
574- `agent.max_concurrent_agents_by_state`: map of positive integers, default `{}`
575- `codex.command`: shell command string, default `codex app-server`
576- `codex.approval_policy`: Codex `AskForApproval` value, default implementation-defined
577- `codex.thread_sandbox`: Codex `SandboxMode` value, default implementation-defined
578- `codex.turn_sandbox_policy`: Codex `SandboxPolicy` value, default implementation-defined
579- `codex.turn_timeout_ms`: integer, default `3600000`
580- `codex.read_timeout_ms`: integer, default `5000`
581- `codex.stall_timeout_ms`: integer, default `300000`
582- `server.port` (extension): integer, optional; enables the optional HTTP server, `0` may be used
583  for ephemeral local bind, and CLI `--port` overrides it
584
585## 7. Orchestration State Machine
586
587The orchestrator is the only component that mutates scheduling state. All worker outcomes are
588reported back to it and converted into explicit state transitions.
589
590### 7.1 Issue Orchestration States
591
592This is not the same as tracker states (`Todo`, `In Progress`, etc.). This is the service's internal
593claim state.
594
5951. `Unclaimed`
596   - Issue is not running and has no retry scheduled.
597
5982. `Claimed`
599   - Orchestrator has reserved the issue to prevent duplicate dispatch.
600   - In practice, claimed issues are either `Running` or `RetryQueued`.
601
6023. `Running`
603   - Worker task exists and the issue is tracked in `running` map.
604
6054. `RetryQueued`
606   - Worker is not running, but a retry timer exists in `retry_attempts`.
607
6085. `Released`
609   - Claim removed because issue is terminal, non-active, missing, or retry path completed without
610     re-dispatch.
611
612Important nuance:
613
614- A successful worker exit does not mean the issue is done forever.
615- The worker may continue through multiple back-to-back coding-agent turns before it exits.
616- After each normal turn completion, the worker re-checks the tracker issue state.
617- If the issue is still in an active state, the worker should start another turn on the same live
618  coding-agent thread in the same workspace, up to `agent.max_turns`.
619- The first turn should use the full rendered task prompt.
620- Continuation turns should send only continuation guidance to the existing thread, not resend the
621  original task prompt that is already present in thread history.
622- Once the worker exits normally, the orchestrator still schedules a short continuation retry
623  (about 1 second) so it can re-check whether the issue remains active and needs another worker
624  session.
625
626### 7.2 Run Attempt Lifecycle
627
628A run attempt transitions through these phases:
629
6301. `PreparingWorkspace`
6312. `BuildingPrompt`
6323. `LaunchingAgentProcess`
6334. `InitializingSession`
6345. `StreamingTurn`
6356. `Finishing`
6367. `Succeeded`
6378. `Failed`
6389. `TimedOut`
63910. `Stalled`
64011. `CanceledByReconciliation`
641
642Distinct terminal reasons are important because retry logic and logs differ.
643
644### 7.3 Transition Triggers
645
646- `Poll Tick`
647  - Reconcile active runs.
648  - Validate config.
649  - Fetch candidate issues.
650  - Dispatch until slots are exhausted.
651
652- `Worker Exit (normal)`
653  - Remove running entry.
654  - Update aggregate runtime totals.
655  - Schedule continuation retry (attempt `1`) after the worker exhausts or finishes its in-process
656    turn loop.
657
658- `Worker Exit (abnormal)`
659  - Remove running entry.
660  - Update aggregate runtime totals.
661  - Schedule exponential-backoff retry.
662
663- `Codex Update Event`
664  - Update live session fields, token counters, and rate limits.
665
666- `Retry Timer Fired`
667  - Re-fetch active candidates and attempt re-dispatch, or release claim if no longer eligible.
668
669- `Reconciliation State Refresh`
670  - Stop runs whose issue states are terminal or no longer active.
671
672- `Stall Timeout`
673  - Kill worker and schedule retry.
674
675### 7.4 Idempotency and Recovery Rules
676
677- The orchestrator serializes state mutations through one authority to avoid duplicate dispatch.
678- `claimed` and `running` checks are required before launching any worker.
679- Reconciliation runs before dispatch on every tick.
680- Restart recovery is tracker-driven and filesystem-driven (no durable orchestrator DB required).
681- Startup terminal cleanup removes stale workspaces for issues already in terminal states.
682
683## 8. Polling, Scheduling, and Reconciliation
684
685### 8.1 Poll Loop
686
687At startup, the service validates config, performs startup cleanup, schedules an immediate tick, and
688then repeats every `polling.interval_ms`.
689
690The effective poll interval should be updated when workflow config changes are re-applied.
691
692Tick sequence:
693
6941. Reconcile running issues.
6952. Run dispatch preflight validation.
6963. Fetch candidate issues from tracker using active states.
6974. Sort issues by dispatch priority.
6985. Dispatch eligible issues while slots remain.
6996. Notify observability/status consumers of state changes.
700
701If per-tick validation fails, dispatch is skipped for that tick, but reconciliation still happens
702first.
703
704### 8.2 Candidate Selection Rules
705
706An issue is dispatch-eligible only if all are true:
707
708- It has `id`, `identifier`, `title`, and `state`.
709- Its state is in `active_states` and not in `terminal_states`.
710- It is not already in `running`.
711- It is not already in `claimed`.
712- Global concurrency slots are available.
713- Per-state concurrency slots are available.
714- Blocker rule for `Todo` state passes:
715  - If the issue state is `Todo`, do not dispatch when any blocker is non-terminal.
716
717Sorting order (stable intent):
718
7191. `priority` ascending (1..4 are preferred; null/unknown sorts last)
7202. `created_at` oldest first
7213. `identifier` lexicographic tie-breaker
722
723### 8.3 Concurrency Control
724
725Global limit:
726
727- `available_slots = max(max_concurrent_agents - running_count, 0)`
728
729Per-state limit:
730
731- `max_concurrent_agents_by_state[state]` if present (state key normalized)
732- otherwise fallback to global limit
733
734The runtime counts issues by their current tracked state in the `running` map.
735
736Optional SSH host limit:
737
738- When `worker.max_concurrent_agents_per_host` is set, each configured SSH host may run at most
739  that many concurrent agents at once.
740- Hosts at that cap are skipped for new dispatch until capacity frees up.
741
742### 8.4 Retry and Backoff
743
744Retry entry creation:
745
746- Cancel any existing retry timer for the same issue.
747- Store `attempt`, `identifier`, `error`, `due_at_ms`, and new timer handle.
748
749Backoff formula:
750
751- Normal continuation retries after a clean worker exit use a short fixed delay of `1000` ms.
752- Failure-driven retries use `delay = min(10000 * 2^(attempt - 1), agent.max_retry_backoff_ms)`.
753- Power is capped by the configured max retry backoff (default `300000` / 5m).
754
755Retry handling behavior:
756
7571. Fetch active candidate issues (not all issues).
7582. Find the specific issue by `issue_id`.
7593. If not found, release claim.
7604. If found and still candidate-eligible:
761   - Dispatch if slots are available.
762   - Otherwise requeue with error `no available orchestrator slots`.
7635. If found but no longer active, release claim.
764
765Note:
766
767- Terminal-state workspace cleanup is handled by startup cleanup and active-run reconciliation
768  (including terminal transitions for currently running issues).
769- Retry handling mainly operates on active candidates and releases claims when the issue is absent,
770  rather than performing terminal cleanup itself.
771
772### 8.5 Active Run Reconciliation
773
774Reconciliation runs every tick and has two parts.
775
776Part A: Stall detection
777
778- For each running issue, compute `elapsed_ms` since:
779  - `last_codex_timestamp` if any event has been seen, else
780  - `started_at`
781- If `elapsed_ms > codex.stall_timeout_ms`, terminate the worker and queue a retry.
782- If `stall_timeout_ms <= 0`, skip stall detection entirely.
783
784Part B: Tracker state refresh
785
786- Fetch current issue states for all running issue IDs.
787- For each running issue:
788  - If tracker state is terminal: terminate worker and clean workspace.
789  - If tracker state is still active: update the in-memory issue snapshot.
790  - If tracker state is neither active nor terminal: terminate worker without workspace cleanup.
791- If state refresh fails, keep workers running and try again on the next tick.
792
793### 8.6 Startup Terminal Workspace Cleanup
794
795When the service starts:
796
7971. Query tracker for issues in terminal states.
7982. For each returned issue identifier, remove the corresponding workspace directory.
7993. If the terminal-issues fetch fails, log a warning and continue startup.
800
801This prevents stale terminal workspaces from accumulating after restarts.
802
803## 9. Workspace Management and Safety
804
805### 9.1 Workspace Layout
806
807Workspace root:
808
809- `workspace.root` (normalized path; the current config layer expands path-like values and preserves
810  bare relative names)
811
812Per-issue workspace path:
813
814- `<workspace.root>/<sanitized_issue_identifier>`
815
816Workspace persistence:
817
818- Workspaces are reused across runs for the same issue.
819- Successful runs do not auto-delete workspaces.
820
821### 9.2 Workspace Creation and Reuse
822
823Input: `issue.identifier`
824
825Algorithm summary:
826
8271. Sanitize identifier to `workspace_key`.
8282. Compute workspace path under workspace root.
8293. Ensure the workspace path exists as a directory.
8304. Mark `created_now=true` only if the directory was created during this call; otherwise
831   `created_now=false`.
8325. If `created_now=true`, run `after_create` hook if configured.
833
834Notes:
835
836- This section does not assume any specific repository/VCS workflow.
837- Workspace preparation beyond directory creation (for example dependency bootstrap, checkout/sync,
838  code generation) is implementation-defined and is typically handled via hooks.
839
840### 9.3 Optional Workspace Population (Implementation-Defined)
841
842The spec does not require any built-in VCS or repository bootstrap behavior.
843
844Implementations may populate or synchronize the workspace using implementation-defined logic and/or
845hooks (for example `after_create` and/or `before_run`).
846
847Failure handling:
848
849- Workspace population/synchronization failures return an error for the current attempt.
850- If failure happens while creating a brand-new workspace, implementations may remove the partially
851  prepared directory.
852- Reused workspaces should not be destructively reset on population failure unless that policy is
853  explicitly chosen and documented.
854
855### 9.4 Workspace Hooks
856
857Supported hooks:
858
859- `hooks.after_create`
860- `hooks.before_run`
861- `hooks.after_run`
862- `hooks.before_remove`
863
864Execution contract:
865
866- Execute in a local shell context appropriate to the host OS, with the workspace directory as
867  `cwd`.
868- On POSIX systems, `sh -lc <script>` (or a stricter equivalent such as `bash -lc <script>`) is a
869  conforming default.
870- Hook timeout uses `hooks.timeout_ms`; default: `60000 ms`.
871- Log hook start, failures, and timeouts.
872
873Failure semantics:
874
875- `after_create` failure or timeout is fatal to workspace creation.
876- `before_run` failure or timeout is fatal to the current run attempt.
877- `after_run` failure or timeout is logged and ignored.
878- `before_remove` failure or timeout is logged and ignored.
879
880### 9.5 Safety Invariants
881
882This is the most important portability constraint.
883
884Invariant 1: Run the coding agent only in the per-issue workspace path.
885
886- Before launching the coding-agent subprocess, validate:
887  - `cwd == workspace_path`
888
889Invariant 2: Workspace path must stay inside workspace root.
890
891- Normalize both paths to absolute.
892- Require `workspace_path` to have `workspace_root` as a prefix directory.
893- Reject any path outside the workspace root.
894
895Invariant 3: Workspace key is sanitized.
896
897- Only `[A-Za-z0-9._-]` allowed in workspace directory names.
898- Replace all other characters with `_`.
899
900## 10. Agent Runner Protocol (Coding Agent Integration)
901
902This section defines the language-neutral contract for integrating a coding agent app-server.
903
904Compatibility profile:
905
906- The normative contract is message ordering, required behaviors, and the logical fields that must
907  be extracted (for example session IDs, completion state, approval handling, and usage/rate-limit
908  telemetry).
909- Exact JSON field names may vary slightly across compatible app-server versions.
910- Implementations should tolerate equivalent payload shapes when they carry the same logical
911  meaning, especially for nested IDs, approval requests, user-input-required signals, and
912  token/rate-limit metadata.
913
914### 10.1 Launch Contract
915
916Subprocess launch parameters:
917
918- Command: `codex.command`
919- Invocation: `bash -lc <codex.command>`
920- Working directory: workspace path
921- Stdout/stderr: separate streams
922- Framing: line-delimited protocol messages on stdout (JSON-RPC-like JSON per line)
923
924Notes:
925
926- The default command is `codex app-server`.
927- Approval policy, cwd, and prompt are expressed in the protocol messages in Section 10.2.
928
929Recommended additional process settings:
930
931- Max line size: 10 MB (for safe buffering)
932
933### 10.2 Session Startup Handshake
934
935Reference: https://developers.openai.com/codex/app-server/
936
937The client must send these protocol messages in order:
938
939Illustrative startup transcript (equivalent payload shapes are acceptable if they preserve the same
940semantics):
941
942```json
943{"id":1,"method":"initialize","params":{"clientInfo":{"name":"symphony","version":"1.0"},"capabilities":{}}}
944{"method":"initialized","params":{}}
945{"id":2,"method":"thread/start","params":{"approvalPolicy":"<implementation-defined>","sandbox":"<implementation-defined>","cwd":"/abs/workspace"}}
946{"id":3,"method":"turn/start","params":{"threadId":"<thread-id>","input":[{"type":"text","text":"<rendered prompt-or-continuation-guidance>"}],"cwd":"/abs/workspace","title":"ABC-123: Example","approvalPolicy":"<implementation-defined>","sandboxPolicy":{"type":"<implementation-defined>"}}}
947```
948
9491. `initialize` request
950   - Params include:
951     - `clientInfo` object (for example `{name, version}`)
952     - `capabilities` object (may be empty)
953   - If the targeted Codex app-server requires capability negotiation for dynamic tools, include the
954     necessary capability flag(s) here.
955   - Wait for response (`read_timeout_ms`)
9562. `initialized` notification
9573. `thread/start` request
958   - Params include:
959     - `approvalPolicy` = implementation-defined session approval policy value
960     - `sandbox` = implementation-defined session sandbox value
961     - `cwd` = absolute workspace path
962     - If optional client-side tools are implemented, include their advertised tool specs using the
963       protocol mechanism supported by the targeted Codex app-server version.
9644. `turn/start` request
965   - Params include:
966     - `threadId`
967     - `input` = single text item containing rendered prompt for the first turn, or continuation
968       guidance for later turns on the same thread
969     - `cwd`
970     - `title` = `<issue.identifier>: <issue.title>`
971     - `approvalPolicy` = implementation-defined turn approval policy value
972     - `sandboxPolicy` = implementation-defined object-form sandbox policy payload when required by
973       the targeted app-server version
974
975Session identifiers:
976
977- Read `thread_id` from `thread/start` result `result.thread.id`
978- Read `turn_id` from each `turn/start` result `result.turn.id`
979- Emit `session_id = "<thread_id>-<turn_id>"`
980- Reuse the same `thread_id` for all continuation turns inside one worker run
981
982### 10.3 Streaming Turn Processing
983
984The client reads line-delimited messages until the turn terminates.
985
986Completion conditions:
987
988- `turn/completed` -> success
989- `turn/failed` -> failure
990- `turn/cancelled` -> failure
991- turn timeout (`turn_timeout_ms`) -> failure
992- subprocess exit -> failure
993
994Continuation processing:
995
996- If the worker decides to continue after a successful turn, it should issue another `turn/start`
997  on the same live `threadId`.
998- The app-server subprocess should remain alive across those continuation turns and be stopped only
999  when the worker run is ending.
1000
1001Line handling requirements:
1002
1003- Read protocol messages from stdout only.
1004- Buffer partial stdout lines until newline arrives.
1005- Attempt JSON parse on complete stdout lines.
1006- Stderr is not part of the protocol stream:
1007  - ignore it or log it as diagnostics
1008  - do not attempt protocol JSON parsing on stderr
1009
1010### 10.4 Emitted Runtime Events (Upstream to Orchestrator)
1011
1012The app-server client emits structured events to the orchestrator callback. Each event should
1013include:
1014
1015- `event` (enum/string)
1016- `timestamp` (UTC timestamp)
1017- `codex_app_server_pid` (if available)
1018- optional `usage` map (token counts)
1019- payload fields as needed
1020
1021Important emitted events may include:
1022
1023- `session_started`
1024- `startup_failed`
1025- `turn_completed`
1026- `turn_failed`
1027- `turn_cancelled`
1028- `turn_ended_with_error`
1029- `turn_input_required`
1030- `approval_auto_approved`
1031- `unsupported_tool_call`
1032- `notification`
1033- `other_message`
1034- `malformed`
1035
1036### 10.5 Approval, Tool Calls, and User Input Policy
1037
1038Approval, sandbox, and user-input behavior is implementation-defined.
1039
1040Policy requirements:
1041
1042- Each implementation should document its chosen approval, sandbox, and operator-confirmation
1043  posture.
1044- Approval requests and user-input-required events must not leave a run stalled indefinitely. An
1045  implementation should either satisfy them, surface them to an operator, auto-resolve them, or
1046  fail the run according to its documented policy.
1047
1048Example high-trust behavior:
1049
1050- Auto-approve command execution approvals for the session.
1051- Auto-approve file-change approvals for the session.
1052- Treat user-input-required turns as hard failure.
1053
1054Unsupported dynamic tool calls:
1055
1056- Supported dynamic tool calls that are explicitly implemented and advertised by the runtime should
1057  be handled according to their extension contract.
1058- If the agent requests a dynamic tool call (`item/tool/call`) that is not supported, return a tool
1059  failure response and continue the session.
1060- This prevents the session from stalling on unsupported tool execution paths.
1061
1062Optional client-side tool extension:
1063
1064- An implementation may expose a limited set of client-side tools to the app-server session.
1065- Current optional standardized tool: `linear_graphql`.
1066- If implemented, supported tools should be advertised to the app-server session during startup
1067  using the protocol mechanism supported by the targeted Codex app-server version.
1068- Unsupported tool names should still return a failure result and continue the session.
1069
1070`linear_graphql` extension contract:
1071
1072- Purpose: execute a raw GraphQL query or mutation against Linear using Symphony's configured
1073  tracker auth for the current session.
1074- Availability: only meaningful when `tracker.kind == "linear"` and valid Linear auth is configured.
1075- Preferred input shape:
1076
1077  ```json
1078  {
1079    "query": "single GraphQL query or mutation document",
1080    "variables": {
1081      "optional": "graphql variables object"
1082    }
1083  }
1084  ```
1085
1086- `query` must be a non-empty string.
1087- `query` must contain exactly one GraphQL operation.
1088- `variables` is optional and, when present, must be a JSON object.
1089- Implementations may additionally accept a raw GraphQL query string as shorthand input.
1090- Execute one GraphQL operation per tool call.
1091- If the provided document contains multiple operations, reject the tool call as invalid input.
1092- `operationName` selection is intentionally out of scope for this extension.
1093- Reuse the configured Linear endpoint and auth from the active Symphony workflow/runtime config; do
1094  not require the coding agent to read raw tokens from disk.
1095- Tool result semantics:
1096  - transport success + no top-level GraphQL `errors` -> `success=true`
1097  - top-level GraphQL `errors` present -> `success=false`, but preserve the GraphQL response body
1098    for debugging
1099  - invalid input, missing auth, or transport failure -> `success=false` with an error payload
1100- Return the GraphQL response or error payload as structured tool output that the model can inspect
1101  in-session.
1102
1103Illustrative responses (equivalent payload shapes are acceptable if they preserve the same outcome):
1104
1105```json
1106{"id":"<approval-id>","result":{"approved":true}}
1107{"id":"<tool-call-id>","result":{"success":false,"error":"unsupported_tool_call"}}
1108```
1109
1110Hard failure on user input requirement:
1111
1112- If the agent requests user input, fail the run attempt immediately.
1113- The client detects this via:
1114  - explicit method (`item/tool/requestUserInput`), or
1115  - turn methods/flags indicating input is required.
1116
1117### 10.6 Timeouts and Error Mapping
1118
1119Timeouts:
1120
1121- `codex.read_timeout_ms`: request/response timeout during startup and sync requests
1122- `codex.turn_timeout_ms`: total turn stream timeout
1123- `codex.stall_timeout_ms`: enforced by orchestrator based on event inactivity
1124
1125Error mapping (recommended normalized categories):
1126
1127- `codex_not_found`
1128- `invalid_workspace_cwd`
1129- `response_timeout`
1130- `turn_timeout`
1131- `port_exit`
1132- `response_error`
1133- `turn_failed`
1134- `turn_cancelled`
1135- `turn_input_required`
1136
1137### 10.7 Agent Runner Contract
1138
1139The `Agent Runner` wraps workspace + prompt + app-server client.
1140
1141Behavior:
1142
11431. Create/reuse workspace for issue.
11442. Build prompt from workflow template.
11453. Start app-server session.
11464. Forward app-server events to orchestrator.
11475. On any error, fail the worker attempt (the orchestrator will retry).
1148
1149Note:
1150
1151- Workspaces are intentionally preserved after successful runs.
1152
1153## 11. Issue Tracker Integration Contract (Linear-Compatible)
1154
1155### 11.1 Required Operations
1156
1157An implementation must support these tracker adapter operations:
1158
11591. `fetch_candidate_issues()`
1160   - Return issues in configured active states for a configured project.
1161
11622. `fetch_issues_by_states(state_names)`
1163   - Used for startup terminal cleanup.
1164
11653. `fetch_issue_states_by_ids(issue_ids)`
1166   - Used for active-run reconciliation.
1167
1168### 11.2 Query Semantics (Linear)
1169
1170Linear-specific requirements for `tracker.kind == "linear"`:
1171
1172- `tracker.kind == "linear"`
1173- GraphQL endpoint (default `https://api.linear.app/graphql`)
1174- Auth token sent in `Authorization` header
1175- `tracker.project_slug` maps to Linear project `slugId`
1176- Candidate issue query filters project using `project: { slugId: { eq: $projectSlug } }`
1177- Issue-state refresh query uses GraphQL issue IDs with variable type `[ID!]`
1178- Pagination required for candidate issues
1179- Page size default: `50`
1180- Network timeout: `30000 ms`
1181
1182Important:
1183
1184- Linear GraphQL schema details can drift. Keep query construction isolated and test the exact query
1185  fields/types required by this specification.
1186
1187A non-Linear implementation may change transport details, but the normalized outputs must match the
1188domain model in Section 4.
1189
1190### 11.3 Normalization Rules
1191
1192Candidate issue normalization should produce fields listed in Section 4.1.1.
1193
1194Additional normalization details:
1195
1196- `labels` -> lowercase strings
1197- `blocked_by` -> derived from inverse relations where relation type is `blocks`
1198- `priority` -> integer only (non-integers become null)
1199- `created_at` and `updated_at` -> parse ISO-8601 timestamps
1200
1201### 11.4 Error Handling Contract
1202
1203Recommended error categories:
1204
1205- `unsupported_tracker_kind`
1206- `missing_tracker_api_key`
1207- `missing_tracker_project_slug`
1208- `linear_api_request` (transport failures)
1209- `linear_api_status` (non-200 HTTP)
1210- `linear_graphql_errors`
1211- `linear_unknown_payload`
1212- `linear_missing_end_cursor` (pagination integrity error)
1213
1214Orchestrator behavior on tracker errors:
1215
1216- Candidate fetch failure: log and skip dispatch for this tick.
1217- Running-state refresh failure: log and keep active workers running.
1218- Startup terminal cleanup failure: log warning and continue startup.
1219
1220### 11.5 Tracker Writes (Important Boundary)
1221
1222Symphony does not require first-class tracker write APIs in the orchestrator.
1223
1224- Ticket mutations (state transitions, comments, PR metadata) are typically handled by the coding
1225  agent using tools defined by the workflow prompt.
1226- The service remains a scheduler/runner and tracker reader.
1227- Workflow-specific success often means "reached the next handoff state" (for example
1228  `Human Review`) rather than tracker terminal state `Done`.
1229- If the optional `linear_graphql` client-side tool extension is implemented, it is still part of
1230  the agent toolchain rather than orchestrator business logic.
1231
1232## 12. Prompt Construction and Context Assembly
1233
1234### 12.1 Inputs
1235
1236Inputs to prompt rendering:
1237
1238- `workflow.prompt_template`
1239- normalized `issue` object
1240- optional `attempt` integer (retry/continuation metadata)
1241
1242### 12.2 Rendering Rules
1243
1244- Render with strict variable checking.
1245- Render with strict filter checking.
1246- Convert issue object keys to strings for template compatibility.
1247- Preserve nested arrays/maps (labels, blockers) so templates can iterate.
1248
1249### 12.3 Retry/Continuation Semantics
1250
1251`attempt` should be passed to the template because the workflow prompt may provide different
1252instructions for:
1253
1254- first run (`attempt` null or absent)
1255- continuation run after a successful prior session
1256- retry after error/timeout/stall
1257
1258### 12.4 Failure Semantics
1259
1260If prompt rendering fails:
1261
1262- Fail the run attempt immediately.
1263- Let the orchestrator treat it like any other worker failure and decide retry behavior.
1264
1265## 13. Logging, Status, and Observability
1266
1267### 13.1 Logging Conventions
1268
1269Required context fields for issue-related logs:
1270
1271- `issue_id`
1272- `issue_identifier`
1273
1274Required context for coding-agent session lifecycle logs:
1275
1276- `session_id`
1277
1278Message formatting requirements:
1279
1280- Use stable `key=value` phrasing.
1281- Include action outcome (`completed`, `failed`, `retrying`, etc.).
1282- Include concise failure reason when present.
1283- Avoid logging large raw payloads unless necessary.
1284
1285### 13.2 Logging Outputs and Sinks
1286
1287The spec does not prescribe where logs must go (stderr, file, remote sink, etc.).
1288
1289Requirements:
1290
1291- Operators must be able to see startup/validation/dispatch failures without attaching a debugger.
1292- Implementations may write to one or more sinks.
1293- If a configured log sink fails, the service should continue running when possible and emit an
1294  operator-visible warning through any remaining sink.
1295
1296### 13.3 Runtime Snapshot / Monitoring Interface (Optional but Recommended)
1297
1298If the implementation exposes a synchronous runtime snapshot (for dashboards or monitoring), it
1299should return:
1300
1301- `running` (list of running session rows)
1302- each running row should include `turn_count`
1303- `retrying` (list of retry queue rows)
1304- `codex_totals`
1305  - `input_tokens`
1306  - `output_tokens`
1307  - `total_tokens`
1308  - `seconds_running` (aggregate runtime seconds as of snapshot time, including active sessions)
1309- `rate_limits` (latest coding-agent rate limit payload, if available)
1310
1311Recommended snapshot error modes:
1312
1313- `timeout`
1314- `unavailable`
1315
1316### 13.4 Optional Human-Readable Status Surface
1317
1318A human-readable status surface (terminal output, dashboard, etc.) is optional and
1319implementation-defined.
1320
1321If present, it should draw from orchestrator state/metrics only and must not be required for
1322correctness.
1323
1324### 13.5 Session Metrics and Token Accounting
1325
1326Token accounting rules:
1327
1328- Agent events may include token counts in multiple payload shapes.
1329- Prefer absolute thread totals when available, such as:
1330  - `thread/tokenUsage/updated` payloads
1331  - `total_token_usage` within token-count wrapper events
1332- Ignore delta-style payloads such as `last_token_usage` for dashboard/API totals.
1333- Extract input/output/total token counts leniently from common field names within the selected
1334  payload.
1335- For absolute totals, track deltas relative to last reported totals to avoid double-counting.
1336- Do not treat generic `usage` maps as cumulative totals unless the event type defines them that
1337  way.
1338- Accumulate aggregate totals in orchestrator state.
1339
1340Runtime accounting:
1341
1342- Runtime should be reported as a live aggregate at snapshot/render time.
1343- Implementations may maintain a cumulative counter for ended sessions and add active-session
1344  elapsed time derived from `running` entries (for example `started_at`) when producing a
1345  snapshot/status view.
1346- Add run duration seconds to the cumulative ended-session runtime when a session ends (normal exit
1347  or cancellation/termination).
1348- Continuous background ticking of runtime totals is not required.
1349
1350Rate-limit tracking:
1351
1352- Track the latest rate-limit payload seen in any agent update.
1353- Any human-readable presentation of rate-limit data is implementation-defined.
1354
1355### 13.6 Humanized Agent Event Summaries (Optional)
1356
1357Humanized summaries of raw agent protocol events are optional.
1358
1359If implemented:
1360
1361- Treat them as observability-only output.
1362- Do not make orchestrator logic depend on humanized strings.
1363

De referentie-implementatie is geschreven in Elixir, want als code in feite gratis is, kun je eindelijk talen kiezen om waar ze goed in zijn, zoals Elixirs sterke ondersteuning voor concurrency. Maar het kernidee past in een eenvoudig Markdown-document. We moedigen je aan om je favoriete programmeeragent op de spec los te laten en een eigen versie te laten implementeren.

De eerste versie van Symphony was gewoon een Codex-sessie die in tmux draaide, Linear voortdurend monitorde en subagents startte voor nieuwe taken. Het werkte, maar was niet bijzonder betrouwbaar. De tweede versie draaide binnen onze hoofdrepo, die vanaf het begin was ingericht voor agents. We hadden de agent-harness al gebouwd om agents de juiste skills en context te geven om in deze repo kwalitatief goed werk te leveren, dus Symphony bracht alles samen.

Toen de basisfunctionaliteit er eenmaal was, gebruikten we Symphony om Symphony te bouwen.

Toen we intern demonstreerden hoe het systeem taken beheerde en de bijbehorende proof-of-work-video toevoegde, was de reactie overweldigend positief: ons Symphony-projectkanaal groeide en teams in de hele organisatie begonnen het spontaan te gebruiken. Sterke interne adoptie is bij OpenAI een voorwaarde om extern te lanceren. Op basis van het gebruik dat we binnen OpenAI zagen, werd duidelijk dat we Symphony ook buiten het bedrijf moesten delen.

Daarom werkten we het idee uit tot een zelfstandige SPEC.md en vroegen we Codex om het te implementeren. Voor de referentie-implementatie kozen we Elixir, een relatief niche programmeertaal met uitstekende bouwstenen voor het orkestreren en aansturen van gelijktijdige processen. Codex bouwde de Elixir-implementatie in één keer en daarna bleven we zowel de spec als de implementatie verder verfijnen. Om de spec nog scherper te krijgen, vroegen we Codex zelfs om die ook in verschillende andere talen te implementeren: TypeScript, Go, Rust, Java en Python. Die resultaten gebruikten we om onduidelijkheden op te sporen en het systeem te vereenvoudigen. Dat lukte in elke taal.

Tijdens het bouwen van Codex hebben we veel onnodige complexiteit verwijderd, zoals afhankelijkheden van specifieke repositories of van Linear MCP. Symphony is niet langer afhankelijk van onze interne repositories of workflows. De kern van de aanpak werd eenvoudig:

Garandeer voor elke open taak dat er een agent draait in zijn eigen werkruimte.

Naast ondersteuning bij het actieve werk is de ontwikkelworkflow nu ook iets wat agents kennen en volgen. Die workflow, zoals werken aan een issue, een repo checken, die op In progress zetten zodat de PM weet dat eraan gewerkt wordt, de PR toevoegen, die naar de status Review verplaatsen en video's toevoegen, is nu vastgelegd in een eenvoudig WORKFLOW.md-bestand. Dit hele proces werd al door mensen gevolgd, maar was nooit gedocumenteerd. In plaats van te vertrouwen op deze impliciete reeks stappen, leggen we die nu vast en zorgt Symphony ervoor dat agents ze volgen. Zo kunnen we agents bouwen die met ons samenwerken. Als we besluiten dat agents ook zelfreflectie aan afgerond werk moeten toevoegen, voegen we dat toe aan de WORKFLOW.md, en Symphony zorgt ervoor dat agents ook die stap volgen.

We konden Codex ook gebruiken in app server mode⁠(opent in een nieuw venster), een ingebouwde headless modus voor Codex. Deze modus stelde ons in staat Codex te draaien en er programmatisch mee te communiceren via een goed gedocumenteerde JSON-RPC API voor dingen zoals het starten van een thread of reageren op turns. Het is een veel handigere en schaalbaardere manier dan proberen met Codex te interageren via de CLI of live tmux-sessies.

Codex App Server paste perfect bij onze use case: we profiteren van de harness die Codex biedt en hebben tegelijk de flexibiliteit en integratiepunten die we nodig hebben. Om bijvoorbeeld te voorkomen dat we het Linear-access-token blootstellen aan subagents, gebruiken we dynamic tool calls⁠(opent in een nieuw venster) om de ruwe functie linear_graphql beschikbaar te maken. Daarmee kunnen willekeurige verzoeken aan Linear worden uitgevoerd, zonder afhankelijk te zijn van MCP of het access-token bloot te stellen aan containers.

Wat hierna volgt

Symphony is een bewust minimale orkestratielaag. We maken het open source om te laten zien wat Codex App Server kan wanneer het wordt gecombineerd met verschillende workflowtools, zoals Linear. Daarom zijn we niet van plan Symphony als zelfstandig product te onderhouden. Zie het als een referentie-implementatie. Net zoals veel developers hun coding agents de post over harness engineering gaven om hun repositories op te zetten, hopen we dat jij je favoriete coding agent de Symphony-spec⁠(opent in een nieuw venster) en repository⁠(opent in een nieuw venster) geeft om je eigen versie te bouwen, afgestemd op jouw omgeving.

De kracht komt van Codex en de app server. Symphony was een manier om Codex met Linear te verbinden, twee tools die we al gebruikten, om het probleem van werkbeheer op te lossen. Naarmate programmeeragents beter worden in redeneren en het volgen van instructies, vermoeden we dat de bottleneck bij andere bedrijven ook verschuift van code schrijven naar het beheren van agent-ondersteund werk. Het mooie is dat de drempel om met dit soort programmeeragentsystemen te experimenteren nu verrassend laag is. Met Codex kun je eenvoudig zelf aan de slag.

Shout-outs aan de community

We zijn verheugd te zien dat de engineeringcommunity Symphony gebruikt in de weken sinds de release, met meer dan 15K GitHub-sterren⁠(opent in een nieuw venster) op 23 april.