Skip to content

gaia-research/gaia-skill-tree / .agents/skills/gaia-steward-retro

gaia-steward-retro

Verified

Produce the Gaia Steward coverage retrospective: the routine table of registered detectors, what each actually observed in the latest receipt, the remaining gaps, and the no-op versus closure record. Covers all 19 routines, including GitHub issue hygiene (18) and routine pull-request maintenance (19). Use for "/gaia-steward-full --retro", "steward coverage report", "which routines are covered", or "how is steward doing".

FileKindSize
SKILL.mdskill md2.0 KB
---
name: gaia-steward-retro
description: >-
  Produce the Gaia Steward coverage retrospective: the routine table of
  registered detectors, what each actually observed in the latest receipt, the
  remaining gaps, and the no-op versus closure record. Covers all 19 routines,
  including GitHub issue hygiene (18) and routine pull-request maintenance (19). Use for
  "/gaia-steward-full --retro", "steward coverage report", "which routines are
  covered", or "how is steward doing".
version: 1.0.0
---

# gaia-steward-retro

## Responsibility

Report coverage honestly: what is registered, what ran, what was blind, and
what the catalog admits it cannot see. For routine 18 the retro also names the
census state of the latest receipt; a stale or missing census is a blind routine,
not a quiet one.

## Input

```bash
gaia steward full run --retro --json
python scripts/check_steward_coverage.py --receipt .gaia/steward/receipts/<run>.json --json
```

## Output

One row per routine (1-19): detectors, live state per detector
(`actively-observed` or `unknown-unavailable`), static states
(`eligible-auto-remediation`, `human-gated`, `partially-covered`), open debt by
class, and `remainingGaps` from `founder/steward/routine-coverage.json`.

## Authority ceiling

Read-only. The retro never edits the catalog; a catalog change is a reviewed
PR that `scripts/check_steward_coverage.py` must accept.

## Stop conditions

- The checker fails: report its errors as the retro's first line and stop.
- No receipt exists: report every detector as `unknown-unavailable`.

## Idempotency

Pure read of the catalog, the policy and one receipt.

## Failure states

Catalog drift (a registered sensor without a contract, a contract without a
sensor, a kind without a binding) fails the checker with exit 1.

## Validation

```bash
python scripts/check_steward_coverage.py
python -m pytest tests/steward/test_routine_coverage.py -q
```

## Tests

`tests/steward/test_routine_coverage.py`, `tests/steward/test_steward_skills.py`.
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 323kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 323kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 323kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall