AgentOps is the operational layer for coding agents.
Publicly, it gives you four things:
- Bookkeeping — captured learnings, findings, and reusable context
- Validation — plan and code review before work ships
- Primitives — single skills, hooks, and CLI surfaces
- Flows — named compositions like
/research,/validate, andrpi
Technically, AgentOps acts as a context compiler: raw session signal becomes reusable knowledge, compiled prevention, and better next work.
Core Flow: RPI
Research → Plan → Implement → Validate
↑ │
└──── Knowledge Flywheel ────┘
Research Phase
/research <topic> # Deep codebase exploration
ao search "<query>" # Search existing knowledge
ao search "<query>" --cite retrieved # Record adoption when a search result is reused
ao lookup <id> # Pull full content of specific learning
ao lookup --query "x" # Search knowledge by relevance
Output: .agents/research/<topic>.md
Plan Phase
/pre-mortem <spec> # Simulate failures (error/rescue map, scope modes, prediction tracking)
/plan <goal> # Decompose into trackable issues
Output: Beads issues with dependencies
Implement Phase
/implement <issue> # Single issue execution
/crank <epic> # Autonomous epic loop (uses swarm for waves)
/swarm # Parallel execution (fresh context per agent)
Output: Code changes, tests, documentation
Validate Phase
/vibe [target] # Code validation (finding classification + suppression + domain checklists)
/post-mortem # Validation + streak tracking + prediction accuracy + retro history
/post-mortem --quick # Quick-capture a single learning (folded the retired retro lane)
Output: .agents/learnings/, .agents/patterns/
Phase-to-Skill Mapping
| Phase | Primary Skill | Supporting Skills |
|---|---|---|
| Discovery | /discovery |
/brainstorm, /research, /plan, /pre-mortem |
| Implement | /crank |
/implement (single issue), /swarm (parallel execution) |
| Validate | /validate |
/vibe, /post-mortem, /forge |
Choosing the skill:
- Use
/implementfor single issue execution. Now defaults to TDD-first — writes failing tests before implementing. Skip with--no-tdd. - Use
/crankfor autonomous epic execution (loops waves via swarm until done). Auto-generates file-ownership maps to prevent worker conflicts. - Use
/discoveryfor the discovery phase only (brainstorm → search → research → plan → pre-mortem). - Use
/validatefor the validation phase only (vibe → post-mortem → forge). - Use
rpifor full lifecycle — delegates to/discovery→/crank→/validate. - Use
/ratchetto gate/record progress through RPI.
Start Here (12 starters)
These are the skills every user needs first. Everything else is available when you need it.
| Skill | Purpose |
|---|---|
/quickstart |
Guided onboarding — run this first |
/bootstrap |
One-command full AgentOps setup — fills gaps only |
/research |
Deep codebase exploration |
/council |
Multi-model consensus review + finding auto-extraction |
/validate |
Canonical PASS/WARN/FAIL verdict over an artifact, plan, code change, PR, or gate |
/vibe |
Code validation (classification + suppression + domain checklists) |
rpi |
Full RPI lifecycle orchestrator (/discovery → /crank → /validate) |
/implement |
Execute single issue |
/post-mortem --quick |
Quick-capture a single learning into the flywheel |
/status |
Single-screen dashboard of current work and suggested next action |
/goals |
Maintain GOALS.yaml fitness specification |
/push |
Atomic test-commit-push workflow |
Advanced Skills (when you need them)
| Skill | Purpose |
|---|---|
/compile, /flywheel |
Active knowledge intelligence and flywheel health — Mine → Grow → Defrag cycle |
/curate |
Canonical miner role for transcripts, .agents/, bd, git, skill diffs, and rare wiki entries |
/curate --mode=harvest |
Cross-rig knowledge consolidation — sweep, dedup, promote to global hub (folded the retired harvest lane) |
/inject |
Operationalize a mature .agents corpus into beliefs, playbooks, briefings, and gap surfaces (folded the retired knowledge-activation lane) |
/brainstorm |
Structured idea exploration before planning |
/discovery |
Full discovery phase orchestrator (brainstorm → search → research → plan → pre-mortem) |
/plan |
Epic decomposition into issues |
/design |
Product validation gate — goal alignment, persona fit, competitive differentiation |
/pre-mortem |
Failure simulation (error/rescue, scope modes, temporal, predictions) |
/post-mortem |
Validation + streak tracking + prediction accuracy + retro history |
/bug-hunt |
Root cause analysis |
/release |
Pre-flight, changelog, version bumps, tag |
/crank |
Autonomous epic loop (uses swarm for each wave) |
/swarm |
Fresh-context parallel execution (Ralph pattern) |
/using-atm |
Run AgentOps loops out of session on an ATM tmux swarm (the ATM leg of the substrate) |
evolve |
Goal-driven fitness-scored improvement loop |
/burndown |
Bounded epic-completion loop — drive a finite target to all-merged, then stop |
/eval-outcomes |
Grade via Outcomes as a holdout-safe projection of the locked eval substrate — one bar, many runtimes |
/operating-loop-workflow |
Install + run the operating-loop multi-agent Workflow (seven-move loop) |
/autodev |
PROGRAM.md autonomous development contract setup and validation |
agy-rules-workflows |
Install AGY-native rules, loop workflow, and scheduled goal controls |
/curate --mode=dream |
Out-of-session knowledge compounding lane |
/doc |
Documentation generation — repo docs (default), gold-standard README (--mode=readme), OSS doc packs (--mode=oss) |
/post-mortem --quick |
Quick-capture a learning (folded the retired retro lane) |
/validate |
Full validation phase orchestrator (vibe → post-mortem → forge) |
/ratchet |
Brownian Ratchet progress gates for RPI workflow |
/forge |
Mine transcripts for knowledge — decisions, learnings, patterns |
/security |
Repository security scanning and release gating, plus the composable binary/prompt-surface suite — static analysis, dynamic tracing, offline redteam, policy gating |
/test |
Test generation, coverage analysis, and TDD workflow |
/cc-hooks |
Author and validate AgentOps runtime hook behavior |
/red-team |
Persona-based adversarial validation — probe docs and skills from constrained user perspectives |
/review |
Review incoming PRs, agent output, or diffs — SCORED checklist |
/refactor |
Safe, verified refactoring with regression testing at each step |
/deps |
Dependency audit, update, vulnerability scanning, and license compliance |
/perf |
Performance profiling, benchmarking, regression detection, and optimization |
/scaffold |
Project scaffolding, component generation, and boilerplate setup |
/scenario |
Author and manage holdout scenarios for behavioral validation |
/skill-auditor |
Two-pass audit of an existing SKILL.md against the unified template (15 checks) |
/skill-builder |
Scaffold or absorb new SKILL.md files against the unified template |
/automation-shape-routing |
Front door for building agent automation — decide the SHAPE (Workflow vs ATM swarm vs plain skill), then hand off to the right builder |
/workflow-builder |
Scaffold a new Claude Workflow script (.claude/workflows/*.js) — deterministic multi-agent orchestration |
/agent-native |
Make out-of-session agents AgentOps-native via skills + ao CLI + CI, not hooks |
Expert Skills (specialized workflows)
| Skill | Purpose |
|---|---|
/swarm |
Parallel Codex agent execution (folded the retired codex-team lane) |
/reverse-engineer-rpi |
Reverse-engineer a product into feature catalog and specs |
/codebase-audit |
Domain-parameterized codebase auditing (security, UX, performance, API, copy, CLI) |
/pr-research |
Upstream repository research before contribution |
/pr-implement |
Fork-based PR implementation |
/validate --mode=pr |
PR-specific validation and isolation checks (folded the retired pr-validate lane) |
/pr-prep |
PR preparation and structured body generation |
/ship-loop |
Bot-paired internal-PR fast-lane cycle |
/complexity |
Code complexity analysis |
/product |
Interactive PRODUCT.md generation |
/handoff |
Session handoff for continuation |
/recover |
Post-compaction context recovery |
/session-bootstrap |
Universal init prompt — every agent runs this first (soc-vuu6.25) |
/trace |
Trace design decisions through history |
/curate --mode=provenance |
Trace artifact lineage to sources |
/beads |
Issue tracking operations |
/heal-skill |
Detect and fix skill hygiene issues |
/converter |
Convert skills to Codex/Cursor formats |
To update installed skills: re-run the install one-liner — bash <(curl -fsSL https://raw.githubusercontent.com/boshu2/agentops/main/scripts/install.sh). (There is no update skill; skill refresh is an install-script concern.)
Knowledge Flywheel
Every /post-mortem promotes learnings and patterns into .agents/ so future /research starts with better context instead of zero.
Inspect, lint, and triage the .agents/ write surface contract via ao agents inspect | lint | doctor (doctor rolls up inspect + lint + orphan/stray-dir report; --strict fails on orphans).
Runtime Modes
AgentOps has several runtime modes. Do not assume hook automation exists everywhere.
| Mode | When it applies | Start path | Closeout path | Guarantees |
|---|---|---|---|---|
substrate (out-of-session) |
A swappable orchestration substrate available out-of-session: an ATM tmux swarm, MCP (ao mcp serve), or managed-agents (ao agent) |
The operator or a lead agent runs bd ready and dispatches a whole loop per bead — an agent that runs the rpi skill; cron / managed triggers run maintenance |
The substrate owns the merge gate (CI-green is the signal) and triggers the knowledge-flywheel feedback | The substrate orchestrates whole rpi/evolve loops (each an agent running the skill) — it never sees the loop's insides; the seam is substrate → agent-running-the-skill. There is no in-CLI runtime=gc executor. See agent-native and docs/3.0.md. |
hook-capable |
Claude/OpenCode with lifecycle hooks installed (no gc) | Runtime hook or ao inject / ao lookup |
Runtime hook or ao forge transcript + ao flywheel close-loop |
Automatic startup/context injection and session-end maintenance when hooks are installed |
codex-native-hooks |
Codex CLI v0.115.0+ with native hook support (March 2026) | Runtime hooks (same as hook-capable) | Runtime hooks (same as hook-capable) | Native lifecycle hooks — same guarantees as hook-capable mode |
codex-hookless-fallback |
Codex Desktop / Codex CLI pre-v0.115.0 without hook surfaces | ao codex start |
ao codex stop |
Explicit startup context, citation tracking, transcript fallback, and close-loop metrics without hooks |
manual |
No hooks and no Codex-native runtime detection | ao inject / ao lookup |
ao forge transcript + ao flywheel close-loop |
Works everywhere, but lifecycle actions are operator-driven |
Issue Tracking
This workflow uses beads for git-native issue tracking:
bd ready # Unblocked issues
bd show <id> # Issue details
bd close <id> # Close issue
bd vc status # Inspect Dolt state if needed (JSONL auto-sync is automatic)
Examples
Startup context loading. AgentOps 3.0's default path is explicit: run ao session bootstrap, then pull prior context with ao inject / ao lookup or a phase-scoped packet. Optional hook-capable runtimes may run an authored session-start.sh, and Codex CLI can use opt-in native hooks via install-codex.sh --with-hooks; those are compatibility/adaptor paths, not the default. Either way the agent gets the RPI workflow, prior context, and a citation path.
Workflow reference during planning. When a user asks how to approach a feature, the agent uses this skill's RPI section to recommend Research → Plan → Implement → Validate — /research for exploration, /plan for decomposition, /pre-mortem for failure simulation — instead of an ad-hoc approach.
Troubleshooting
| Problem | Cause | Solution |
|---|---|---|
| Skill not loaded | Startup path not run | Run ao session bootstrap, then ao inject / ao lookup; for Codex lifecycle recovery, run ao codex start explicitly |
| Outdated skill catalog | This file not synced with actual skills/ directory | Update skill list in this file after adding/removing skills |
| Wrong skill suggested | Natural language trigger ambiguous | User explicitly calls skill with /skill-name syntax |
| Workflow unclear | RPI phases not well-documented here | Read full workflow guide in README.md or docs/ARCHITECTURE.md |