Governed control plane and learning layer above Claude Code: a gated task pipeline with signed human approval gates and a gating switch (running-control-plane), project setup and signing-identity guidance (setting-up-agentic-os), and eval-gated skill improvement in an isolated lab (improving-skills-in-lab).
Not in a marketplace yet, so it loads for one session with --plugin-dir.
Plugin agent-agentic-os in richfrem/agent-plugins-skills. Commands read from the repo's manifest — yesterday.
improving-skills-in-labRuns the Skill Improvement Lab in the main session: plan, bootstrap an isolated lab repo, KEEP or DISCARD improvement loop with a locked scorer, log and chart runs, backport, close with memory. Use for "improve this skill", "evaluate this skill", "score this change", "KEEP or DISCARD", "skill is not
running-control-planeCRITICAL INTAKE GATEWAY: use at the start of any non-trivial engineering task, feature, refactor, multi-file bugfix, work package or new idea, including requests to explore, brainstorm, build, fix or change something, before planning or code changes. Enforces read-only discovery, 1-3 structured ques
setting-up-agentic-osSets up and retrofits a project for the Agentic OS and guides a human through creating the approval signing key. Use for "set up agentic OS", "initialize agent harness", "init my project for AI agents", "retrofit repository", "upgrade project for evolution", "sync instruction files", "where do I put
Developer utilities: 8 self-contained skills. Manages GitHub issues (log, promote, rank, refactor proposals, resolution audits) and resolves them in isolated worktrees through PR closure; bundles code and docs for review; checks links; enforces coding conventions; manages ADRs, symlinks, and HuggingFace dataset/model transfers.
8 skills
Ecosystem authoring primitives — creates and audits agent skills, plugins, sub-agents, instructions, rules, hooks, GitHub workflows, and APM packages.
7 skills
Business exploration and reengineering: guided discovery, requirements capture, interactive prototypes, risk-tiered handoff to engineering, and rescue of existing vibe-coded prototypes. Phase gates enforced by running-exploration-cycle.
5 skills
Composable agent loop and graph architectures — deterministic state machines, DAG workflows, supervisor/worker loops, and parallel swarms, paired with the multi-LLM task router suite (run_agent.py, 13 expert agent personas, AGT sandboxing, and MAF validation). Framework-agnostic: works with any CLI agent.
5 skills
Agent Memory & Knowledge: filesystem memory, RLM summary ledgers, vector retrieval, and Obsidian vault and wiki workflows.
4 skills
Turn any domain folder of skills into a bounded agentic loop: a manifest builder inventories a domain's skills/tools/checks, a goal compiler turns a goal into a verifiable task plan (refusing vague goals with forcing questions), and a JSON-backed loop controller drives execute->verify->close with retry caps, controller-run verification (no verification theater), human escalation on exhausted budgets, and a close gate that refuses while any task is unverified or unwaived. Ships 3 stdlib Python tools, 18 committed per-domain harness manifests + JSON schema, 3 references citing the 2024-2026 agent-harness canon (Anthropic long-running harnesses, verifier's law, SWE-agent, Ralph loop, Cognition), harness-runner agent + /cs-harness command. Use when an agent or subagent should pick up a goal, define its tasks, complete and verify them, and close the loop.
1 skill · 1 command · 1 agent
Build an AI-ready engineering system for safe coding-agent delivery and continuous software improvement.
11 skills
Accountable, time-bounded Codex shifts with a persistent punch list, parked decisions, and reviewable run history.
16 skills
Ten skills that test your agent harness on your own machine: sandbox reach, guardrails, 'tests pass' claims, rules the agent breaks, token waste, regressions after updates, MCP tool design, and a test drive of coding agents on tasks from your own git history.
10 skills
Intent to implementation: ready, then go.
3 skills
Keep long Claude Code runs on task: enforce the punch list, park questions, recover crashes, run product or quality shifts until done, and leave a receipt for every item with its tokens and time.
16 skills · hooks
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.