project-toolkit plugin
Complete project development toolkit: agents, reusable skills, and lifecycle skills (/spec to /ship) for Claude Code workflows
How to install the project-toolkit plugin
Run both commands inside Claude Code. The first adds the marketplace once; the second installs the plugin.
/plugin marketplace add rjmurillo/ai-agents
/plugin install project-toolkit@ai-agentsWhat's inside project-toolkit
135 parts read from the plugin's own files.
Agents 29
Call as @agent-project-toolkit:name- @agent-project-toolkit:analyst
No description in the file.
- @agent-project-toolkit:architect
No description in the file.
- @agent-project-toolkit:code-reviewer
No description in the file.
- @agent-project-toolkit:code-simplifier
No description in the file.
- @agent-project-toolkit:comment-analyzer
No description in the file.
- @agent-project-toolkit:critic
No description in the file.
- @agent-project-toolkit:debug
No description in the file.
- @agent-project-toolkit:dependency-auditor
No description in the file.
- @agent-project-toolkit:devops
No description in the file.
- @agent-project-toolkit:explainer
No description in the file.
- @agent-project-toolkit:high-level-advisor
No description in the file.
- @agent-project-toolkit:implementer
No description in the file.
- @agent-project-toolkit:independent-thinker
No description in the file.
- @agent-project-toolkit:issue-feature-review
No description in the file.
- @agent-project-toolkit:janitor
No description in the file.
- @agent-project-toolkit:merge-resolver
No description in the file.
- @agent-project-toolkit:milestone-planner
No description in the file.
- @agent-project-toolkit:negotiation
No description in the file.
- @agent-project-toolkit:orchestrator
No description in the file.
- @agent-project-toolkit:pr-test-analyzer
No description in the file.
- @agent-project-toolkit:qa
No description in the file.
- @agent-project-toolkit:quality-auditor
No description in the file.
- @agent-project-toolkit:retrospective
No description in the file.
- @agent-project-toolkit:roadmap
No description in the file.
- @agent-project-toolkit:security
No description in the file.
- @agent-project-toolkit:silent-failure-hunter
No description in the file.
- @agent-project-toolkit:skillbook
No description in the file.
- @agent-project-toolkit:task-decomposer
No description in the file.
- @agent-project-toolkit:type-design-analyzer
No description in the file.
Skills 105
Claude loads it when the task matches- adr-generator
Generates structured Architectural Decision Records (ADRs) by researching existing templates, gathering context, and determining the next sequential number.
- adr-review
Orchestrates a multi-agent debate to validate Architecture Decision Records through specialized roles like architect, critic, and security.
- agent-harness-reference
Provides technical specifications for Claude Code and GitHub Copilot CLI hook contracts, including payloads, exit codes, and event behaviors.
- ai-agents-external-claims
Verifies external, vendor, and third-party claims against authoritative primary sources to prevent inaccurate data from entering repository artifacts.
- analysis-provenance
Identifies code ownership by analyzing file headers and package manifests to prevent accidental modification of upstream dependencies.
- analyze
Performs systematic multi-step codebase analysis including architecture reviews, security assessments, and code quality evaluations.
- api-doc-authoring
Write developer-centric API documentation using use cases, executed examples, and failure recovery remediation.
- autoplan
Routes requests to the appropriate skill, command, or agent chain by classifying intent and task size.
- avoiding-manufactured-work
Detects and classifies unnecessary follow-up work to ensure agents only produce deliverables demanded by real consumers or acceptance criteria.
- benchmark-models
Compares LLM providers like Claude, GPT, and Gemini side-by-side on latency, cost, tokens, and quality using a standardized benchmark script.
- build
No description in the file.
- business-strategy
Routes founder business problems to specific decision-tree frameworks distilled from 14 influential business books.
- buy-vs-build-framework
Provides a strategic framework for evaluating build, buy, partner, or defer decisions using TCO analysis and a four-phase process.
- chaos-experiment
Design and document rigorous chaos engineering experiments to validate system resilience through failure injection and hypothesis testing.
- checkpoint
Creates timestamped, secret-redacted snapshots of decisions and progress and links them to the active session log.
- chestertons-fence
Investigates the historical context, original rationale, and dependencies of code or architectural decisions before changes are proposed.
- code-qualities-assessment
Assess code maintainability by scoring cohesion, coupling, encapsulation, testability, and non-redundancy using quantifiable rubrics.
- codebase-documenter
Scaffold structured project documentation templates including README, ARCHITECTURE, and API references with bracketed placeholders.
- codeql-scan
Execute static security analysis scans with automated language detection, database caching, and SARIF output generation.
- context-gather
Gathers multi-source context from memory, documentation, and web searches to provide a synthesized summary before planning or implementation.
- context-hub-setup
Checks for Serena and Context7 plugin prerequisites and provides installation commands for any missing components.
- context-optimizer
Optimizes agent context by classifying content placement, compressing markdown to pipe-delimited format, and validating compliance with decision frameworks.
- curating-memories
Maintains markdown memory files by marking superseded content with strikethroughs and dated banners to keep current truth visible.
- cva-analysis
Discovers natural software abstractions by building a matrix of commonalities and variabilities across requirements to inform design pattern selection.
- cynefin-classifier
Classifies problems into specific domains to recommend appropriate cognitive approaches and response strategies.
- decision-critic
Stress-tests reasoning by surfacing hidden assumptions, verifying claims, and generating adversarial perspectives to improve decision quality.
- decision-question
Enables an agent to present concise decision briefs to users only when they are blocked by decisions the user owns.
- destructive-change-safety
Verifies the blast radius and recovery plans for destructive operations like deletions, overwrites, or resource drops before execution.
- doc-accuracy
Verifies documentation accuracy by treating code as the source of truth through multi-phase audits of claims, compilability, and consistency.
- domain-design
Designs APIs, schemas, and module boundaries based on business behavior and invariants rather than source payload shapes or generic CRUD.
- dx-review
Performs evidence-based developer experience audits by testing onboarding, API ergonomics, documentation, and error messages to produce a scored scorecard.
- evidence-provenance
Classifies the lineage and independence of evidence to determine if claims are corroborated by multiple reasoning loops or a single loop.
- execution-plans
Manage execution plans as versioned artifacts with progress tracking, decision logs, and lifecycle management.
- fix-markdown-fences
Scan and repair malformed markdown code fence closings to prevent content bleeding and unclosed blocks.
- git-advanced-workflows
Performs advanced Git operations including rebasing, cherry-picking, bisecting, worktree management, and commit recovery using reflog.
- github
Execute GitHub operations like managing pull requests, issues, milestones, and CI status using Python scripts with structured output.
- github-url-intercept
Routes GitHub URLs to efficient API calls instead of fetching heavy HTML to optimize context window usage.
- golden-principles
Scans repositories for violations of specific governance principles and provides agent-readable remediation instructions.
- maintainer-review
Evaluates pull requests and issues to provide a disposition decision on whether a change should exist and what action a maintainer should take.
- memory
Routes requests to specific sub-skills within a tiered memory system for searching, recording, maintaining, or consolidating information.
+65
Hooks 1
Runs on these events- hooks
No description in the file.
Using project-toolkit in Claude Code
Ask for an agent by name:
@agent-project-toolkit:analyst
@agent-project-toolkit:architectSkills need no command. Claude reads them when your request matches what they describe.
Update, turn off or remove
- Update the plugin (shell)
- claude plugin update project-toolkit@ai-agents
- Refresh the marketplace
- /plugin marketplace update ai-agents
- Turn it off for now
- /plugin disable project-toolkit@ai-agents
- Remove it
- /plugin uninstall project-toolkit@ai-agents
Skills in project-toolkit
- analyzerjmurillo/ai-agents
Performs systematic multi-step codebase analysis including architecture reviews, security assessments, and code quality evaluations.
11 files49stars - benchmark-modelsrjmurillo/ai-agents
Compares LLM providers like Claude, GPT, and Gemini side-by-side on latency, cost, tokens, and quality using a standardized benchmark script.
2 files49stars - checkpointrjmurillo/ai-agents
Creates timestamped, secret-redacted snapshots of decisions and progress and links them to the active session log.
2 files49stars - codebase-documenterrjmurillo/ai-agents
Scaffold structured project documentation templates including README, ARCHITECTURE, and API references with bracketed placeholders.
3 files49stars - context-optimizerrjmurillo/ai-agents
Optimizes agent context by classifying content placement, compressing markdown to pipe-delimited format, and validating compliance with decision frameworks.
22 files49stars - pr-autofixrjmurillo/ai-agents
Automatically triages, fixes CI failures, addresses thread feedback, and enables auto-merge for pull requests based on defined readiness tiers.
4 files49stars - pr-reviewrjmurillo/ai-agents
Respond to pull request review comments and run a completion gate to determine if a PR is ready to merge.
3 files49stars - requirements-interviewrjmurillo/ai-agents
Conducts adversarial requirements interviews or audits existing specifications to elicit testable requirements and resolve ambiguities.
7 files49stars - researchrjmurillo/ai-agents
Performs deep external research to produce long-form analysis documents, Serena memories, and drafted GitHub issues.
3 files49stars - reviewrjmurillo/ai-agents
Performs multi-stage code reviews by evaluating spec-compliance and risk-selected technical axes like security, architecture, and reliability.
27 files49stars - shiprjmurillo/ai-agents
Runs pre-flight validation checks and creates a pull request or validates a contributor branch for GitHub or Azure DevOps.
3 files49stars - dx-reviewrjmurillo/ai-agents
Performs evidence-based developer experience audits by testing onboarding, API ergonomics, documentation, and error messages to produce a scored scorecard.
49stars - planrjmurillo/ai-agents
Decomposes a technical specification into milestones, atomic tasks, dependency graphs, and risk registers.
49stars - pr-quality-allrjmurillo/ai-agents
Runs six specialized quality axes against working changes and merges their verdicts into a single consolidated report.
49stars - pr-quality-analystrjmurillo/ai-agents
Assesses local code diffs for quality, blast radius, and maintainability to return a PASS, WARN, or CRITICAL_FAIL verdict.
49stars - pr-quality-architectrjmurillo/ai-agents
Reviews local git diffs for design patterns, system boundaries, coupling, cohesion, and ADR compliance to provide a quality verdict.
49stars - pr-quality-devopsrjmurillo/ai-agents
Evaluates local git diffs for CI/CD, build pipeline, and infrastructure risks to return a PASS, WARN, or CRITICAL_FAIL verdict.
49stars - pr-quality-qarjmurillo/ai-agents
Evaluates local code diffs against specific QA criteria to provide a PASS, WARN, or CRITICAL_FAIL verdict on test coverage and error handling.
49stars - pr-quality-roadmaprjmurillo/ai-agents
Assess local git diffs for strategic alignment, scope creep, and user value to return a PASS, WARN, or CRITICAL_FAIL verdict.
49stars - pr-quality-securityrjmurillo/ai-agents
Scans local git diffs for vulnerabilities, secret exposure, and OWASP Top 10 anti-patterns to provide a security verdict.
49stars - push-prrjmurillo/ai-agents
Commits changes, pushes branches, and opens pull requests using repository templates populated from the actual git diff.
49stars - retrorjmurillo/ai-agents
Fills unfilled auto-retrospective skeletons for a specific date or lists all pending retrospective skeletons.
49stars - adr-generatorrjmurillo/ai-agents
Generates structured Architectural Decision Records (ADRs) by researching existing templates, gathering context, and determining the next sequential number.
6 files49stars - adr-reviewrjmurillo/ai-agents
Orchestrates a multi-agent debate to validate Architecture Decision Records through specialized roles like architect, critic, and security.
14 files49stars - agent-harness-referencerjmurillo/ai-agents
Provides technical specifications for Claude Code and GitHub Copilot CLI hook contracts, including payloads, exit codes, and event behaviors.
3 files49stars - ai-agents-external-claimsrjmurillo/ai-agents
Verifies external, vendor, and third-party claims against authoritative primary sources to prevent inaccurate data from entering repository artifacts.
2 files49stars - analysis-provenancerjmurillo/ai-agents
Identifies code ownership by analyzing file headers and package manifests to prevent accidental modification of upstream dependencies.
2 files49stars - api-doc-authoringrjmurillo/ai-agents
Write developer-centric API documentation using use cases, executed examples, and failure recovery remediation.
6 files49stars - autoplanrjmurillo/ai-agents
Routes requests to the appropriate skill, command, or agent chain by classifying intent and task size.
10 files49stars - business-strategyrjmurillo/ai-agents
Routes founder business problems to specific decision-tree frameworks distilled from 14 influential business books.
15 files49stars
project-toolkit plugin questions
How do I install the project-toolkit plugin?
In Claude Code run /plugin marketplace add rjmurillo/ai-agents, then /plugin install project-toolkit@ai-agents.
What does the project-toolkit plugin include?
project-toolkit bundles 29 agents, 105 skills, hooks.
Which agents come with project-toolkit?
analyst, architect, code-reviewer, code-simplifier, comment-analyzer, critic +23. Call one with @agent-project-toolkit:analyst.
How do I update or remove the project-toolkit plugin?
Update it with claude plugin update project-toolkit@ai-agents in your shell, or from the Installed tab of /plugin. Remove it with /plugin uninstall project-toolkit@ai-agents.