evaluator
claudeEvaluator-optimizer pattern: adversarial review, verification gates (TaskCompleted), static analysis, postmortems, reference-fidelity check, self-evolution proposal assessment.
14 skills · 1 agent · hooks
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.
Plugin context-engine 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.
audit-contextUse when sessions feel slow, CLAUDE.md has grown unwieldy, MCP server overhead is unclear, or you want to audit token overhead before a large task — measures CLAUDE.md size, system reminder weight, MCP server token cost, and per-skill description length, then ranks the top offenders.
belief-auditUse when a long session has compacted or handed off, before any Edit, to confirm Claude's belief about file contents still matches disk. Diffs current sha256 signatures against the last-recorded snapshot in `.claude/state/belief.jsonl` for the N most-recently-edited files (default 5).
checkpointUse during long sessions when the conversation feels like it may have drifted, when an unrelated subgoal has appeared, or when tracking-error matters before the next major edit — runs a fast comparison of recent work against the original task statement, listing scope creep, context bloat, and unfulf
context-tricksUse when the user wants to learn techniques for long or complex sessions — covers guided compaction, partial compaction, side-question dispatch, checkpointing, and session discipline. Reference-style passive skill, applied inline whenever the conversation hits context-management questions.
lean-mdUse when a CLAUDE.md file has grown past 100 lines, when compliance has dropped (the model is ignoring rules), or when the user asks to "trim" / "shorten" / "audit" their CLAUDE.md — applies the every-line-must-earn-its-place principle and rewrites the file with only the directives that change obser
token-pipelineRun the 5-stage Token Transformation Pipeline (TRAE §5.2.2) over the current session state — Collection, Ranking, Compression, Budgeting, Assembly. Emits a concrete next-action recommendation (/compact, /lean-md, or /progress-log + fresh session) instead of a generic warning.
W=${CLAUDE_PLUGIN_ROOT}/diagnostics-lib/time-hook.sh; H=${CLAUDE_PLUGIN_ROOT}/hooks/env-bootstrap.sh; if [ -x "$W" ]; then bash "$W" context-engine SessionStart bash "$H"; else bash "$H"; fi
W=${CLAUDE_PLUGIN_ROOT}/diagnostics-lib/time-hook.sh; H=${CLAUDE_PLUGIN_ROOT}/hooks/mcp-instruction-monitor.sh; if [ -x "$W" ]; then bash "$W" context-engine SessionStart bash "$H"; else bash "$H"; fi
bash ${CLAUDE_PLUGIN_ROOT}/hooks/track-context-pressure.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/track-system-reminders.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/task-guardian.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/task-guardian-log.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/check-large-file.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/warn-tool-truncation.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/track-edits.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/detect-thrashing.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/belief-verify.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/plan-mode-enter.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/consecutive-failure-reset.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/consecutive-failure-guard.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/pre-compact-guard.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/pre-compact.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/forward-briefing.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/post-compact.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/post-compact-belief-audit.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/post-compact-recovery.sh
bash ${CLAUDE_PLUGIN_ROOT}/hooks/belief-snapshot.sh
Evaluator-optimizer pattern: adversarial review, verification gates (TaskCompleted), static analysis, postmortems, reference-fidelity check, self-evolution proposal assessment.
14 skills · 1 agent · hooks
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
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
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
Self-evolution boundary: change-manifest writer, evolution-history ledger, session-digest, auto-tune-skill outer loop, controllability invariant (POLICY.md).
8 skills · hooks
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
Appends two or three short follow-up questions after a substantive answer to help users check key facts, probe the reasoning, and notice missing context
19 skills
Tenant-isolated AI teams that plan work, coordinate tasks, recall approved context, and return evidence-backed results
6 skills · 6 commands · 4 agents · 1 MCP server
Skill: add headroom-ai compress() to a Python or TypeScript app to cut LLM input tokens (tool outputs, logs, JSON, chat history). No API key.
1 skill
Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.
1 MCP server
Perform AI code review on Git diffs — supports workspace changes, branch ranges, and single commits with concurrent per-file analysis, codebase search, and deep context-aware review.
2 commands
Your code review partner. CodeRabbit provides external validation using a specialized AI architecture and 40+ integrated static analyzers—offering a different perspective that catches bugs, security vulnerabilities, logic errors, and edge cases. Context-aware analysis via AST parsing and codegraph relationships. Automatically incorporates CLAUDE.md and project coding guidelines into reviews. Useful after writing or modifying code, before commits, when implementing complex or security-sensitive logic, or when a second opinion would increase confidence in the changes. Returns specific findings with suggested fixes that can be applied immediately. Free to use.
2 skills · 1 command · 1 agent
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.