| name | simba-codex-lifecycle |
| description | Enforce Simba's Codex lifecycle routine for coding tasks. Use when starting or finishing implementation work in a Simba-enabled repo to run `simba codex-status` at start, handle any still-pending raw Codex transcript extraction, and run `simba codex-finalize` before final handoff. |
Simba Codex Lifecycle
Run this routine for implementation tasks in Simba-enabled repositories.
Required Sequence
- Recall relevant memory first (use the user request as query text):
simba codex-recall "<user request>"
- Run status once per session:
simba codex-status
Do not run this on every prompt. Re-run only if:
- there was a long pause or major context shift,
- memory actions fail unexpectedly,
- you are about to finalize and want a fresh check.
- Codex
PreCompact is the primary automatic session-analysis path. If
codex-status still reports a raw JSONL transcript as pending_extraction
and you need to analyze it now, run the explicit storage path:
simba codex-extract --run
If that reports no candidates or fails, run the manual prompt path:
simba codex-extract
- Before final handoff, run finalize:
simba codex-finalize
Finalize Inputs
Prefer to include both response text and transcript path when available:
simba codex-finalize --response-file /path/to/response.txt --transcript /path/to/transcript.jsonl
If those files are not available, still run simba codex-finalize as a best-effort check.
If simba is not in PATH, use:
uvx --from /Users/mahmoud/src/ai/simba simba codex-status
Output Discipline
- Mention whether
codex-recall returned relevant memories and how they informed your approach.
- Mention in your final update that lifecycle checks were run.
- If you run
codex-extract --run, mention the stored / duplicate / error
counts. If extraction remains pending, run codex-extract and report that
action.
