| name | decisions-codex-worker |
| description | Use when Codex is running work that originated from DecisionsAI tickets, projects, workflows, or Initiative proposals. |
DecisionsAI Codex Worker
When a task comes from DecisionsAI:
- Treat the instruction as project-bound work. Use the provided ticket, board,
workflow, and project context as the source of truth.
- Keep edits scoped to the project folder and the explicit ticket outcome.
- If the prompt starts with result-packet or previous-step context, use that as
state handed over by the DecisionsAI workflow. Do not discard it when
implementing or validating the current step.
- Respect the model selected by DecisionsAI/Codex runtime. Do not claim a
different model was used unless the runtime explicitly reports it.
- Report status in a shape DecisionsAI can write back:
Status: completed | failed | needs_input
Summary: ...
Files changed: ...
Tests: ...
Evidence: ...
Blockers: ...
Next step: ...
- If a ticket lacks enough context, return
Status: needs_input with the
specific missing decision instead of inventing requirements.
- If the task is part of a workflow, include any artifact paths or validation
output the next workflow step needs.
- If DecisionsAI asks for a practical implementation, make a real, inspectable
artifact rather than a plan-only response unless the instruction is explicitly
an audit or planning step.
- Prefer the Codex IDE/chat surface as the primary execution context. The CLI
is fallback transport for automation or setup checks, not the canonical
conversation. Keep responses structured so DecisionsAI can checkpoint,
retry, escalate, continue, or close the workflow step.
- When a task is too broad for one pass, complete the safest useful slice and
set
Next step to the exact continuation DecisionsAI should queue.
- Apply the DecisionsAI harness prerequisites on every implementation task:
- decisions-harness-stack — master index (ECC, Ponytail, Fallow, RTK, browser/content).
- Ponytail (
ponytail skill) — YAGNI ladder, stdlib-first, minimal diff; mark shortcuts with ponytail: comments.
- Fallow (
fallow skill) — on JS/TS repos, run npx fallow audit --format json --quiet || true before reporting complete and attach the verdict in Tests: or Evidence:.
- Browser / QA —
browser-qa, decisions-playwright, webapp-testing when touching UI; use Hermes playwright_browser when Decisions is running.
- Content / video —
content-engine, article-writing, remotion-video-creation, fal-ai-media when the ticket is content work (see ~/.decisions/harness/mcp-recommendations.json for fal MCP).
- UI ideation —
decisions-ui-ideation then decisions-design-references before new screens; use Refero/Mobbin MCP when configured (see ~/.decisions/harness/mcp-setup-design.sh).
- Internet research —
decisions-agent-reach + agent-reach for URLs, social, video, GitHub, RSS; run agent-reach doctor --json first.
- 30-day synthesis —
last30days when the ticket needs “what people said lately” (heavier than agent-reach; needs last30days scripts + optional API keys).
- Publishable copy —
humanizer after drafts; decisions-marketing-skills + product-marketing for GTM work.
- YouTube / subtitles —
decisions-yt-dlp or workflow ytdlp steps; not for Bilibili (use agent-reach bili).
- SaaS apps (Gmail, Slack, Notion, Jira) —
decisions-composio (Composio Connect MCP); not Rube (deprecated).
- For any normal Codex IDE/chat prompt inside a DecisionsAI project folder,
even when it did not originate from a workflow packet, report the turn to
DecisionsAI as a project IDE session. At the start of the prompt, call:
python3 ~/plugins/decisions-codex/scripts/report_decisions_event.py --event-type codex_prompt_submitted --status observed --input "<user prompt>" --thread-id "<codex thread id when known>"
Before the final response, call:
python3 ~/plugins/decisions-codex/scripts/report_decisions_event.py --event-type codex_completed --status completed --output "<short result summary>" --thread-id "<codex thread id when known>"
If both sides of a completed turn are only available at the end, call:
python3 ~/plugins/decisions-codex/scripts/report_decisions_event.py --turn-input "<user prompt>" --turn-output "<short result summary>" --thread-id "<codex thread id when known>"
The reporter resolves the DecisionsAI project from the current working
directory and creates or resumes the Codex IDE session. If DecisionsAI is
not running, the reporter exits quietly and Codex continues normally.
- If the prompt includes a
[DECISIONS CODEX CALLBACK] block, treat that as
live workflow metadata. Use the callback URL or reporter script in that
block to report meaningful events back to DecisionsAI:
codex_started when work begins.codex_prompt_submitted after every user prompt or Codex instruction
submission connected to this DecisionsAI run, including follow-up prompts
typed directly in Codex.user_steer when the human changes direction, adds constraints, or
corrects the approach.codex_waiting or codex_needs_input when blocked on a decision.codex_interrupted when the current task is paused or superseded.codex_progress for material implementation milestones.codex_completed or codex_failed when the work finishes.
If there is no workflow callback block, still report project chat events
through the reporter script. It will use the current working directory to
create or resume a DecisionsAI project IDE session:
python3 ~/plugins/decisions-codex/scripts/report_decisions_event.py --event-type codex_prompt_submitted --status observed --message "<what changed>" --thread-id "<codex thread id when known>"
- Do not keep steering only inside the Codex conversation. If the human gives
new direction while a DecisionsAI workflow is running, report it so Hermes
can store the event, update workflow memory, and let the orchestrator decide
whether to continue, validate, retry, or ask a follow-up.
- Before reporting an event, assume DecisionsAI may or may not be open. If
DecisionsAI is reachable and the callback or reporter succeeds, continue
normally. If DecisionsAI is not reachable, keep working without surfacing a
bridge error to the user unless the task explicitly asks for diagnostics.
- If Codex is working in a fresh chat without a DecisionsAI workflow callback,
still report material starts, progress, steering, blockers, and completion.
Inside a project folder, use the IDE session reporting in item 12. For
activity that is not attached to a known project workflow, use ambient
harness reporting with the current project folder:
python3 ~/plugins/decisions-codex/scripts/report_decisions_event.py --source ambient --harness codex --event-type codex_progress --status observed --message "<what changed>"
Ambient reporting is best-effort. If DecisionsAI is switched off or
unreachable, the reporter must fail silently and Codex work continues.
