agent-builder
Create custom Claude Code sub-agents with specialized expertise, limited tools, and reusable prompts.
brain
Search project knowledge - code, docs, expertise, patterns
conductor
Prompt optimization - scores, analysis, variant generation
create-failing-test
Create failing test specs for TDD: document reproduction steps and assertions for Dev/QA handoff.
create-qa-task
Generate QA task spec from validated issue: scenarios, test data, and handoff package for QA agent.
document-project
Use to analyze and document any project codebase. Creates comprehensive reference documentation for AI-assisted development including architecture and patterns.
execute-checklist
Validate documents, stories, or processes against defined checklists with pass/fail report.
file-first
Read source files directly with Glob/Grep/Read. No RAG, no vector databases, no pre-loaded summaries.
hooks-manager
Create, configure, test, and debug Claude Code hooks for lifecycle automation across 9 event types.
init-context
Initialize PRISM .context folder in a new project with context modules and CLAUDE.md setup.
initialize-architecture
Use to create complete architecture documentation structure. Creates all required architecture documents from templates.
investigate-root-cause
Systematically investigate bugs to find root cause via code analysis, git history, and error tracing.
test-design
Design E2E integration test strategy for a story: scenarios, patterns, and test spec document.
create-dev-task
Generate a dev task document for the Dev agent based on validated issue and investigation results.
apply-qa-fixes
Apply code/test fixes from QA gate results: prioritized plan from gate YAML and assessments.
correct-course
Handle sprint changes and scope adjustments: analyze impact, re-estimate stories, and adjust sprint capacity.
shared
Shared reference documentation: PRISM best practices, command patterns, and cross-skill examples.
nfr-assess
Use to assess non-functional requirements (security, performance, reliability, maintainability) through E2E integration testing patterns.
qa-gate
Create/update QA gate YAML for a story: PASS/CONCERNS/FAIL/WAIVED decision with actionable issues.
remember
Persist patterns, conventions, decisions, or failures to Mulch expertise for cross-session memory.
risk-profile
Use to assess and document risk factors for stories or features. Creates risk profiles with mitigation strategies.
sdlc-handoff
Use for SDLC phase transitions. Ensures proper handoff between development phases with documentation.
skill-builder
Build, optimize, and validate Claude Code skills using progressive disclosure and token efficiency.
strangler-pattern
Safely migrate legacy controllers using strangler pattern with feature flags
trace-requirements
Map acceptance criteria to E2E integration tests: build traceability matrix and identify coverage gaps.
validate-issue
Validate reported issues using Playwright: confirm reproducibility and document evidence for Dev team.
validate
Run PRISM quality checks - docs, links, and portability. Use before committing or when something looks broken.
version-bump
>