rptc

Plugin rptc from marketplace rptc-workflow-marketplace in kuklaph/rptc-workflow. Commands read from the repo's manifest — 16 days ago.

Skills (27)

• architect-methodology

  Plan a design when interfaces, data shapes, ownership, sequencing, or migration are genuinely uncertain, including high-risk work. Skip changes that follow an established pattern, even across several files.

• brainstorming

  Resolve genuine product intent, requirements, and design choices through focused dialogue. Use when the repository and runtime cannot answer a consequential question. Skip discoverable facts and reversible implementation details.

• code-review-methodology

  Review a change against the request, concrete correctness risks, and the repository's documented standards. Use inside RPTC flows or when the user asks for an RPTC review of a diff, branch, or pull request.

• core-principles

  RPTC's portable engineering invariants. Load for RPTC feature, fix, review, and verification work.

• diagnose-methodology

  Reproduce and diagnose a reported bug, regression, flake, crash, or performance problem whose cause is not proven. Use when behavior is broken, intermittent, or slow. Skip when an existing failing check already isolates a known mechanical fix.

• docs-methodology

  Determine documentation required by changed public behavior, APIs, configuration, migrations, or operating procedures. Use for RPTC documentation review.

• frontend-design

  Aesthetic direction for new user-facing UI or explicit redesign and polish requests. Not for bug fixes or small edits inside an existing design system.

• html-report-generator

  Produce a saved report as one self-contained HTML file in a consistent house style. Use when the user asks for an HTML report or a saved report deliverable.

• research-methodology

  Investigate codebase behavior, external facts, or both using sources appropriate to the claim. Use inside RPTC flows or when the user asks for a sourced investigation of unfamiliar APIs, standards, current practice, or a repository compared with external guidance. Skip ordinary lookups.

• rptc-codex-init

  Install or refresh the packaged RPTC Codex custom-agent TOML files in the project-local or explicitly requested global Codex agents directory.

• rptc-commit

  Run project-defined checks, stage selected paths, and create only the commit or pull request explicitly requested through the RPTC ship contract.

• rptc-config

  Discover the project's check commands (tests, typecheck, lint, build) and propose a short Checks section for AGENTS.md.

• rptc-feat

  Implement a feature with RPTC, scaling research, planning, TDD, delegation, and verification to the change's uncertainty and risk. Use for RPTC feature or refactoring requests.

• rptc-fix

  Reproduce, diagnose, fix, and verify a bug with RPTC. Use for broken, failing, flaky, crashing, or unexpectedly slow behavior whose cause is not proven.

• rptc-research

  Research a codebase, external question, or hybrid comparison using evidence appropriate to the claim and clear fact/inference separation.

• rptc-structure

  Audit concrete codebase structure friction and propose evidence-backed refactors that improve module depth, ownership, dependency direction, locality, or testability.

• rptc-test-impact

  Audit changed behavior against tests and independent contracts without assuming implementation or tests are automatically correct. Use for failing assertions after behavior changes, missing regression protection, orphaned tests, and unclear test ownership.

• rptc-verify-loop

  Fix accepted RPTC verification findings and recheck the affected evidence until claims are verified, failed, inconclusive, or explicitly left open. Use for iterative verify-fix-recheck work.

• rptc-verify

  Verify RPTC acceptance claims, changed risks, and repository fit using direct checks and selected report-only agents. Use for an RPTC verification pass on a diff, path, or codebase.

• security-methodology

  Review changed security properties at trust boundaries, including authorization, untrusted input, secret handling, dependencies, sensitive data, and failure behavior. Use for security-sensitive diffs and RPTC verification.

• structure-methodology

  Assess or design codebase structure using module depth, interface size, ownership, dependency direction, locality, and testability. Use for explicit structure audits and non-trivial architecture work.

• tdd-agent-methodology

  Execution contract for an RPTC implementation agent using vertical TDD. Use when a delegated implementation agent owns a bounded code change.

• tdd-methodology

  Build changed behavior through small failing-then-passing vertical slices at a stable seam. Use when behavior changes and a practical executable test path exists. Skip for prototypes, pure documentation, and cases where a new test would be brittle or disproportionately expensive.

• test-impact-methodology

  Audit how changed behavior affects tests using requirements, public contracts, prior verified behavior, tests, and current implementation as ordered evidence. Use for failing assertions after behavior changes, missing regression protection, orphaned tests, and unclear test ownership.

• unslop-writing-clearly

  Edit prose deliverables (documentation, reports, commit messages, PR descriptions, release notes, user-facing copy) to remove AI writing patterns while keeping the intended voice. Not for code-only work.

• verification-evidence

  Classify completion claims as VERIFIED, NOT VERIFIED, or INCONCLUSIVE and attach the command, artifact, or runtime observation that supports each status. Use during RPTC verification and completion reporting.

• agent-teams

  Claude-only orchestration for several independent substantial workstreams or a deliberate multi-perspective debate. Use explicitly when work can be divided into exclusive ownership or independent artifacts.

Similar plugins