Simba Codex Lifecycle

Simba Codex Lifecycle

Run this routine for implementation tasks in Simba-enabled repositories.

Required Sequence

1. Recall relevant memory first (use the user request as query text):

simba codex-recall "<user request>"

2. 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.

3. If status reports pending_extraction, run:

simba codex-extract

4. 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 codex-status reports pending extraction, do not ignore it. Run codex-extract and report that action.

Converted and distributed by TomeVault — claim your Tome and manage your conversions.