aia-harness

Plugin aia-harness from marketplace leandro-plugins-registry in leandrosilvaferreira/claude-plugins-registry. Commands read from the repo's manifest — today.

Skills (8)

• check-consistency-workflow

  Whole-harness consistency audit for an existing Claude Code project: enumerates every skill, agent, rule, command, script, and CLAUDE.md file, cross-references every mention against what actually exists on disk, dispatches parallel auditors to settle uncertain references and judge whether each artif

• condense-harness-prompts-workflow

  Condenses harness .md prompts (.claude/agents, commands, rules, skills) using caveman full + Opus, without losing information. Compresses verbose prose and tables, preserves code blocks/URLs/inline-code/headings via a deterministic gate. Overwrites in place (git is the review safety net). Use when a

• harness-engineering

  Use when setting up, scaffolding, bootstrapping, or improving a Claude Code "harness" for a project — hooks, skills, agents, rules, settings.json, .mcp.json, settings.local.json, worktree config, or per-domain CLAUDE.md. Triggers on "set up Claude Code", "harness engineering", "scaffold hooks/rules"

• mcp-catalog

  Use when choosing, recommending, or wiring MCP servers for a project's .mcp.json. Curated list of strategic MCP servers with safe configuration (env placeholders, scoping). Triggers on "add an MCP", "which MCP servers", "configure .mcp.json", "strategic MCPs".

• revise-agent-frontmatter

  Review and upgrade the routing descriptions of candidate agents under templates/ (ECC, ag-kit, first-party) so each follows the best-practice standard — condition-shaped, with "Use proactively" and explicit trigger conditions. Writes the canonical description into the matching *_AGENT_WHEN_TO_USE ma

• revise-agent-routing-workflow

  Audit .claude/agents/*.md frontmatter descriptions against the routing-description standard (condition-shaped, "Use proactively", 40-600 chars) and sync them with the project's CLAUDE.md `## Workflow & Agents` table or free-text mentions. Fixes descriptions in place, fixes stale/missing table rows,

• revise-claude-md

  Generate or refresh rich intermediate CLAUDE.md files for strategic subdirectories of the target project. Reads .claude/rules/ (recursive), analyzes actual source files, maps rules to domains, and generates domain-specific CLAUDE.md with Key patterns, Applied rules, and Local conventions sections. T

• safe-hooks

  Use when writing, wiring, reviewing, or debugging Claude Code hooks — especially JavaScript hooks. Covers exit codes, the node-resolver wrapper, fail-open vs blocking behavior, matchers, timeouts, and security. Triggers on "write a hook", "PreToolUse/PostToolUse/Stop hook", "hook not blocking", "for

Slash commands (15)

• /aia-harness:add-github-pm

  Add GitHub PM pillar to an existing harness

• /aia-harness:add-mcp

  Suggest and generate strategic MCP server entries for the project's .mcp.json (env placeholders only).

• /aia-harness:add-obsidian

  Install and configure the Obsidian vault-as-memory pillar (vault skeleton, MCP server, hooks, rule) into this project

• /aia-harness:add-plugins

  Install the suggested Claude Code plugins for this stack (code-review, hookify, feature-dev, frontend-design, context7, claude-code-setup + per-language LSP). One confirmation, then auto-installs.

• /aia-harness:add-tools

  Install token-economy / code-graph tools. Caveman and ponytail install as global Claude Code plugins; rtk hook + claude-code-worktrees are vendored project-level automatically; binary/pkg installs run after one confirmation.

• /aia-harness:audit-prompts

  Review CLAUDE.md files, rules, skills, agents and commands for prompt text written for older Claude models and propose a diff fitted to the models the project pins — nothing is written until you approve.

• /aia-harness:check-consistency

  Audit an existing harness end to end — every cross-reference between skills, agents, rules, commands, scripts and CLAUDE.md files, plus whether each artifact still fits the project's real stack — then fix what you approve.

• /aia-harness:check-deps

  Check required system dependencies (Node, Python, Go, etc.) before harness operations.

• /aia-harness:condense-harness-prompts

  Validate/fix Claude Code artifact frontmatters and semantically condense harness .md files (.claude/agents, commands, rules, skills) with Opus + deterministic gate.

• /aia-harness:doctor

  Audit an existing Claude Code harness (CLAUDE.md quality, settings safety, hook hygiene) and propose targeted fixes.

• /aia-harness:help

  Show the full aia-harness command reference — what each command does, when to use it, parameters and options — with a "I want to…" quick-start guide at the top.

• /aia-harness:init

  Scan the project and scaffold a full Claude Code harness — diagnose, approve, then apply with diffs.

• /aia-harness:patch

  Merge the current harness into a project that already has one — apply everything that is safe mechanically, then adjudicate every conflicting file with the user, diff by diff.

• /aia-harness:revise-agent-routing

  Audit agent frontmatter descriptions in .claude/agents and sync them with CLAUDE.md routing mentions on an existing target project.

• /aia-harness:scan

  Scan the current project and print a Claude Code harness diagnosis (read-only, writes nothing).

Subagents (4)

• architecture-mapper

  Maps a codebase into architectural domains/layers so per-domain CLAUDE.md files can be generated with accurate responsibilities. Use during harness init for non-trivial or monorepo projects. Read-only.

• harness-consistency-auditor

  Audits one category of an existing Claude Code harness (skills, agents, rules, commands+scripts, or CLAUDE.md files) for cross-reference breakage a deterministic pass could not settle, and for content that no longer fits the project's real stack. Read-only; returns severity-tagged findings, never ed

• harness-reviewer

  Adversarially audits a freshly scaffolded Claude Code harness for safety and quality before it is trusted — secrets, fail-open hooks, over-broad permissions, bloated CLAUDE.md. Use right after applying a harness plan.

• stack-analyst

  Deep-dives a project's stack when the deterministic scan needs human-level judgment — ambiguous frameworks, unusual build setups, polyglot repos, or missing canonical commands. Read-only; returns a concise structured summary.

Hooks (1)

• SessionStart

  node

Similar plugins