claude

forge-meta

Self-evolution boundary: change-manifest writer, evolution-history ledger, session-digest, auto-tune-skill outer loop, controllability invariant (POLICY.md).

3 on GitHubv1.0.0updated 5 months agoself-evolution
/

Plugin forge-meta from marketplace forge-studio in ArtMin96/forge-studio. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (8)

  • auto-tune-skill

    Outer-loop proposer that iterates on a single SKILL.md body, scores each candidate via /run-evals-bench, computes a Pareto frontier on (pass_rate, token_cost), and writes the Pareto-best revision to .claude/proposals/<plugin>-<skill>-<timestamp>.md for human review. Never modifies the original SKILL

  • change-manifest

    Write a structured change-manifest entry to `.claude/evolution/change_manifest.jsonl`. Each entry captures what changed, why, predicted impact, what was read, what assumptions were made, and how to verify the work — forming the evolution ledger that `forge-meta` reads for session digests and history

  • evolution-history

    Use when you want to review the project's evolution ledger — renders the append-only change manifest as a reverse-chronological Markdown timeline grouped by date (newest first), capped at the last 200 entries.

  • harness-metrics

    Compute seven harness-level quality dimensions (arXiv:2605.18747 §5.2.1, arXiv:2605.26112 §4.2) from existing Forge Studio artifacts and render a Markdown scorecard. Use when you want a snapshot of how well the harness is performing across trajectory efficiency, verification strength, recovery abili

  • manifest-analyze

    Use when you want a structured report on what the change manifest contains — frequency of failure patterns, risk task distribution, constraint-level split, and top "why this component" clusters. Reads `.claude/evolution/change_manifest.jsonl` and emits a 5-section Markdown report to stdout.

  • paper-research

    Use when the user names a paper PDF in `docs/papers/` and asks to research, audit, or apply its findings to this marketplace — produces a single research brief that grounds the paper's claims in the existing plugin/hook/skill/agent surface, prefers reuse before proposing new, and verifies every Clau

  • session-digest

    Produce a ≤10KB Markdown digest of the current session's evolution artifacts. Writes to `.claude/sessions/<session-id>-digest.md`. Useful after any multi-agent run to get a compact rollup without reading raw JSONL files.

  • skill-staleness-audit

    Score every SKILL.md in the marketplace against staleness signals (last-edit age, eval coverage, citation freshness, frontmatter compliance) and emit a ranked report. Read-only — never modifies a skill.

Hooks (3)

  • PreToolUse

    bash ${CLAUDE_PLUGIN_ROOT}/hooks/pre-edit-guard.sh

  • SubagentStop

    bash ${CLAUDE_PLUGIN_ROOT}/hooks/manifest-writer.sh

  • SessionEnd

    bash ${CLAUDE_PLUGIN_ROOT}/hooks/session-end-digest.sh

View source on GitHub

More from ArtMin96/forge-studio

All 21 plugins →

evaluator

claude

Evaluator-optimizer pattern: adversarial review, verification gates (TaskCompleted), static analysis, postmortems, reference-fidelity check, self-evolution proposal assessment.

14 skills · 1 agent · hooks

3

workflow

claude

Hook-driven agentic workflow orchestrator. Automatic routing (shell or LLM-hybrid), sprint-contract enforcement, TDD loop, session bootstrap, handoff nudges. Self-evolution SEPL loop (propose/assess/commit/rollback) over versioned resources. Composes agents/evaluator/context-engine/memory rather than duplicating them.

11 skills · hooks

3

agents

claude

Multi-agent decomposition: planner/generator/reviewer pipeline triad with tool-isolated capability boundaries, plus a read-only researcher for pre-planning investigation. Includes /worktree-team for physical worktree isolation, directory-ownership enforcement (opt-in), and output-schema checks on generator exit.

5 skills · 4 agents · hooks

3

diagnostics

claude

Codebase health scanning: documentation drift, registration gaps, convention violations, stale memory, HARNESS_SPEC invariant compliance. /validate-marketplace performs pre-commit mechanical correctness checks (distinct from /entropy-scan). /startup-profile reports SessionStart hook latency. /policies-list prints the cross-plugin enforcement registry (FS01–FS42).

8 skills · hooks

3

traces

claude

Execution trace collection, analysis, and harness evolution. Stores structured JSONL traces for pattern detection and failure-driven improvement. Feeds /evolve as a proposal source, plus per-session clarification-waste analysis.

7 skills · hooks

3

context-engine

claude

Context management: progressive pressure tracking, drift detection, edit safety, large-file warnings, plan mode hooks, MCP monitoring, prompt budget awareness, failure-escalation with safe-mode flag.

6 skills · hooks

3

Similar plugins

CMEM Pro

claude

CMEM Pro for the Claude app - hooks stream tool use to cmem.ai and inject observations into new sessions and agents

3 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

ruflo-observability

claude

Structured logging, distributed tracing, and metrics — correlate agent swarm activity with application telemetry

2 skills · 1 command · 1 agent

ruvnet/ruflo 81,589

agentic-bundle-aas-observability-ir

claudecodex

Install the "AAS Observability & Incident Response" editorial skill bundle for Claude Code.

12 skills

sickn33/agentic-awesome-skills 49,620

honeycomb

claudecodex

Skills, agents, and workflows for Honeycomb observability — query patterns, production investigations, SLOs, OpenTelemetry instrumentation, and Beeline migration. Designed to complement the Honeycomb MCP server.

11 skills · 1 command · 2 agents · 1 MCP server · hooks

anthropics/claude-plugins-official 47,076

kubernetes-operations

claudecodex

Kubernetes manifest generation, networking configuration, security policies, observability setup, GitOps workflows, and auto-scaling

4 skills · 1 agent

wshobson/agents 42,461

Honeycomb

claudecodex

Skills, agents, and workflows for Honeycomb observability — query patterns, production investigations, SLOs, OpenTelemetry instrumentation, and Beeline migration. Designed to complement the Honeycomb MCP server.

11 skills · 1 command · 2 agents · 1 MCP server · hooks

anthropics/knowledge-work-plugins 35,601

Maintain forge-meta?

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.

claude & codex plugins21 plugins