## Docs site — boshu2.github.io/agentops
- Replace Jekyll minima with MkDocs Material (slate dark palette default)
- Landing hero scoped to .landing-hero (no gradient bleed on inner pages)
- Section landings (getting-started, architecture, reference, contracts, levels) with narrative ledes
- Generate skill catalog + per-skill pages from skills/*/SKILL.md at build
- Generate CLI reference from cli/docs/ at build (fails build on missing source)
- Fix every internal link via one-shot migration (scripts/migrations/2026-04-19-fix-docs-links.py)
- Wire mkdocs build --strict into pre-push gate (Check 25a)
- Replace .github/workflows/pages.yml with .github/workflows/docs.yml (actions/deploy-pages)
- README + docs/CONTRIBUTING updated with docs-build toolchain on-ramp
- tests/docs/validate-links.sh patched to understand mkdocs-generated paths
## RPI orchestrator strict-delegation contract
- Strict sub-skill delegation is now the default for /rpi, /discovery, /validation. No --full flag — always on.
- Canonical contract in skills/shared/references/strict-delegation-contract.md (plus Codex mirror)
- Forged learning with detection phrases + corrective actions (ao-inject-ready)
- Rename --no-lifecycle -> --no-scaffold in /discovery (Claude + Codex) with deprecation alias through v2.40.0
- Harden inline-work hotspots with orchestrator-owned markers in /crank STEPS 3a.3, 6.5, 8.7 + Inline Work Policy footer
- Flag consistency: --auto and --interactive scopes documented; --complexity syntax formalized
- CHANGELOG [Unreleased] documents + scaffolds; runtime hook enforcement explicitly deferred
## fix(hooks): go-test-precommit skips when no Go files staged
- Previously ran go vet + go build unconditionally, blocking non-Go commits when Go toolchain was missing or at a different version
- Now checks staged diff and exits cleanly if no *.go or go.{mod,sum,work} files are staged
- Applied to both hooks/ (source) and cli/embedded/hooks/ (synced)
Provenance: full RPI lifecycle with strict delegation at every phase boundary (Phase 1 Skill(discovery) -> Phase 2 Skill(crank) -> Phase 3 Skill(validation) -> Skill(vibe), Skill(post-mortem), Skill(retro)).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Thin orchestrator that spawns context-restricted agents to probe whether
docs and skills actually work when someone with constraints tries to use
them. Complements /council (expert judgment) and /vibe (code quality)
with consumer-side usability testing.
- Add skills/red-team/ with SKILL.md (336L) and 4 reference docs
- Add red-team preset to council personas (4 personas: OnCall, NewHire, Agent, Consumer)
- Add /red-team to using-agentops catalog and SKILL-TIERS.md
- Sync skill counts to 65 across all docs
- stability=experimental, v1 surfaces: docs + skills only
Codex gpt-5.4 reviewed discovery plan; all 5 critiques addressed in v2.
New `ao harvest` command walks all .agents/ directories across the
workspace, extracts learnings/patterns/research, deduplicates by
content hash, and promotes high-confidence items to the global hub.
- cli/internal/harvest: walker, extract, catalog packages (25 tests)
- cli/cmd/ao/harvest.go: cobra command with --roots, --promote-to,
--min-confidence, --dry-run, --json flags
- skills/harvest: /harvest skill with governance reference
- skills-codex/harvest: codex parity
- config: HarvestRoots field in PathsConfig
- docs: skill counts synced to 61 (52 user-facing)
Fix right-side │ border alignment in box diagrams. Three fix types:
- Trailing space adjustment for border alignment
- Box border widening for content that exceeded width
- Bottom border dash count matching top border
Implements the first three issues of the Athena Knowledge Intelligence Engine:
- ao mine: extracts knowledge signal from git history, .agents/research/
orphans, and code complexity hotspots. Outputs JSON to .agents/mine/.
Supports --sources, --since (26h/7d/etc), --output, --quiet, --dry-run.
- ao defrag: prunes orphan learnings, deduplicates near-duplicates via
trigram overlap, sweeps oscillating evolve goals from cycle-history.jsonl.
Outputs JSON to .agents/defrag/. All operations non-destructive by default.
- /athena skill: orchestrates Mine → Grow → Defrag cycle with LLM-driven
validation of existing learnings against current code, cross-domain
synthesis, gap identification, and report generation.
Also: adds mine+defrag to cobra command registration test, syncs skill
counts from 50→51 across all docs, updates embedded skill copy.
Pre-mortem fixes applied: SinceSeconds as int64 (not time.Duration),
co-change algorithm clarified as files-in-≥3-commits sorted by frequency.
Wave 3 of ag-lhb: Remove deleted skill entries from SKILL-TIERS.md,
sync counts to 49 total (40 user-facing, 9 internal), fix broken
GLOSSARY.md link to deleted extract skill, fix heal.sh dead xrefs.
- CHANGELOG.md: comprehensive v2.18.0 section
- AGENTS.md, README.md, docs/: sync doc updates for new commands/features
- skills/converter: enhance Codex conversion script
- tests/hooks/test-hooks.sh: add test cases for new hook features
Phase 1 — index at startup: ao inject --index-only renders a compact
table (~200 tokens) instead of full content (~800 tokens). Enabled via
AGENTOPS_INDEX_INJECT=1 feature flag in session-start.sh.
Phase 2 — on-demand pull: ao lookup <id> fetches full content by ID,
ao lookup --query "topic" searches by relevance, ao lookup --bead ag-xyz
filters by source bead lineage. Records "pulled" citations for MemRL.
Also: remove 3 unused test helpers (writeGoalsMD, setupRatchetChain,
setupAgentsDirs) to fix lint, add prune policies for evolve/vibe/crank/
brainstorm/local-ci directories, update skills and docs.
Agent Mail was an experimental coordination layer for tmux-based
distributed workers. Local mode (native teams) covers all production
use cases. Removing to simplify the codebase before evolving remote
dispatch via bushido.
Deleted: inbox skill, agent-mail-protocol, 3 distributed-mode
reference docs, 2 test files. Edited 15 files to remove references.
Update README, PRODUCT.md, ARCHITECTURE.md, and how-it-works.md to
reflect production learnings from the 116-cycle overnight evolve run:
- Two-tier execution model: orchestrators stay visible, workers fork
- Compaction-proof disk-backed state with hard-gated verification
- Evolve example updated with real overnight metrics
- PRODUCT.md strengthened with production validation data
- Add Custom Agents section to ARCHITECTURE.md explaining agentops:researcher
and agentops:code-reviewer purpose vs Explore and general-purpose agents
- Add collapsible custom agents section to README.md
- Sync embedded copies of enhanced go.md, common-standards.md, and SKILL.md
Boden Fuller