| 2026-10-02 15:06 |
v1.8.0 — Removed R4 (skill-accuracy) — skill evaluation moved to skill-eval. agentfs-eval now focuses on workspace health only. |
| 2026-10-02 12:35 |
v1.7.1 — Fixed A3 EXPECTED_RULES 8→7: Rule 8 (Path Hygiene) was removed in v7.0.0 but check script was not updated |
| 2026-09-29 21:27 |
v7.2.0 — v7.2.0 — rules consolidated 10→8, severity tags removed |
| 2026-09-29 14:44 |
v1.7.0 — Updated A3 and A8 assertions for v7.0.0 heading-format rules |
| 2026-09-27 23:47 |
v1.6.2 — Tighten behavioral tests: expected first tool is search_nodes only (not search_nodes OR load_skill), load_skill moved to anti-pattern list |
| 2026-09-27 23:35 |
v1.6.1 — Fix B2 expected tool order: search_nodes before load_skill, matching Rule 2 fallthrough sequence |
| 2026-09-27 20:18 |
v1.6.0 — Update template-check.sh for agentfs v6.0.0: A2 removes Scopes requirement (Rules must be 1st section), A3 expects 10 rules (not 17), A5 checks template-wide for search_nodes/load_skill (not a single rule row), adds A2c Signal Dispatch ordering |
| 2026-09-25 01:46 |
v1.5.7 — Update template-check.sh: 17 rules, A5 checks Rule 7 instead of Rule 8 |
| 2026-09-25 00:22 |
v1.5.6 — B1 test: use ‘check headroom status’ (no CLI-name collision); document command-shape ambiguity |
| 2026-09-24 22:50 |
v1.5.5 — Resilient session cleanup: trap EXIT handler, per-session cleanup fn, orphan sweep on updated_at, DB timeout=5s |
| 2026-09-24 22:13 |
v1.5.4 — Exit 0 always + anti-loop warning — prevents weaker models from re-running eval in a loop |
| 2026-09-24 21:14 |
v1.5.3 — Update assertion docs: A2 Scopes before Rules, A4 context lookup fallback, A5 search_nodes + load_skill, remove stale A6 |
| 2026-09-24 19:24 |
v1.5.2 — SQLite-based test session cleanup instead of goose session remove |
| 2026-09-24 17:25 |
v1.5.1 — Fix double-run bug: capture output from single run instead of running checks twice |
| 2026-09-24 17:17 |
v1.5.0 — template-eval.sh reads AGENT_SESSION_ID to detect provider/model from current session. Zero args required. Fallback chain: session metadata → config.yaml |
| 2026-09-24 16:51 |
v1.4.0 — template-eval.sh accepts –provider/–model args, falls back to config.yaml. Quick Dispatch passes session provider/model. |
| 2026-09-24 16:05 |
v1.3.0 — Rename agents-md-* scripts to template-*, add template-eval.sh zero-arg wrapper, add merge-score-entry.sh, simplify Quick Dispatch |
| 2026-09-24 13:03 |
v1.2.0 — Add agents-md-check.sh: deterministic AGENTS.md template quality assertions (A1-A9) |
| 2026-08-26 21:34 |
v1.1.0 — Added LITE scope awareness: skips skills/profiles L1 checks for LITE projects. |
| 2026-07-13 15:38 |
v1.0 — Initial design: three-layer eval (structural, behavioral, semantic), maturity levels L0–L5, explicit trigger only, fresh session recommendation |