qrspi

Plugin qrspi in lotea-be/qrspi. Commands read from the repo's manifest — 1 month ago.

Skills (20)

• qrspi-dogfood

  How to dogfood an in-flight QRSPI change in THIS repo before the PR stage — load the local working copy as the plugin with `claude --plugin-dir`, build throwaway consumer fixtures, and walk the change's `(human)` verification tasks against a live session so their tasks.md boxes become observed Confi

• qrspi-marketplace-bump

  Bump the qrspi plugin's pinned source ref in the separate lotea-be/ai-agent-marketplace repo to a published release tag, as a reviewed PR (never a direct push to the marketplace main). Locates the marketplace repo as a sibling folder, verifies the target tag is actually published, edits only the qrs

• qrspi-release

  Cut a tag-based release of the QRSPI kit. Bumps plugin.json, rolls CHANGELOG [Unreleased] into a dated version section, re-checks lint, commits, and pushes main — then creates the tag and, after an explicit human confirmation, pushes it to publish (which triggers release.yml). The skill never pushes

• qrspi-stack

  Stack cheatsheet for qrspi -- languages, runtime versions, frameworks, key libraries, project layout, testing, and coding conventions. This is the QRSPI stack-cheatsheet skill for this repo; load it whenever you need the project's tech stack or conventions.

• backlog-writer

  Shared append procedure for adding a Check-22-valid idea row to openspec/backlog.md. Load this skill whenever you need to capture a separable future change as a new idea row. Handles dedup, P-band proposal, row construction, and staging.

• context-budget-gate

  In-session context-budget gate for QRSPI orchestrators. Tracks a stage-event counter in conversational context, fires a one-line advisory nudge at 8 events (no AskUserQuestion), and a never-suppressed soft-gate AskUserQuestion at 12 events. Load this at the top of each gate-scoped QRSPI command, aft

• context-hygiene

  Rules for keeping agent context windows lean (target under 40%, reset at 60%) and using subagents as context firewalls rather than personas. Load this when planning a multi-stage workflow, starting a long session, or deciding whether to delegate work to a subagent.

• git-host-workflow

  Shared derivation procedure for git-host and remote awareness -- resolves the host CLI (gh/az repos/glab), PR-create/PR-status commands, and branch-naming slots (feature/archive) that /qrspi:pr, /qrspi:archive, and /qrspi:questions need. Load this skill whenever a command needs to resolve a vendor C

• implementer-core

  Full shared implementer body for the QRSPI Implement stage. Carries all preconditions, coding rules, what-to-do steps, stuck-handling, and output format. Load this from the effort-variant agents (implementer-low, implementer-medium, implementer-high) to keep a single source of truth for implementer

• openspec-archive-change

  Archive a completed change in the experimental workflow. Use when the user wants to finalize and archive a change after implementation is complete.

• openspec-sync-specs

  Sync delta specs from a change to main specs. Use when the user wants to update main specs with changes from a delta spec, without archiving the change.

• openspec-workflow

  How this project uses OpenSpec to persist QRSPI artifacts. Covers folder layout and how each QRSPI stage maps onto OpenSpec artifacts. Load this whenever you need to create, update, or archive a change folder under openspec/.

• postpr-fix

  How to make a transient, qrspi-aware code fix AFTER the PR has been opened — the small follow-ups that surface from the reviewer's "Open issues" list or a retrospective code flag. Keeps code, tests, and the change's DELTA spec in sync, ticks the follow-up, and commits atomically on the PR branch. Lo

• qrspi-update

  Drives /qrspi:update. Carries the migration-manifest schema contract, the SemVer-ordered walk algorithm (apply each migrations/<v>.yaml for A < v <= B in ascending order), target-version resolution (auto-detect primary, explicit-arg fallback), the edge-case handling (up-to-date / no-marker / downgra

• qrspi-version-check

  Session-scoped version check for QRSPI commands. Compares the repo's installed-kit marker (A from openspec/.qrspi-version) against the installed plugin version (B from Claude Code's installed_plugins.json registry under $CLAUDE_CONFIG_DIR) and branches on the result -- silent on match, AskUserQuesti

• repo-surface

  Shared kit skill that owns the surface taxonomy, section-to-surface mapping, and omit/warn rules for all five artifact-producing QRSPI agents. Load this in each agent's preamble to determine which sections to emit based on the project's stack cheatsheet.

• retrospective

  How to run a post-stage retrospective on a just-completed QRSPI stage. Captures friction the agent encountered that the prompts/skills did not anticipate, and turns it into concrete edits to the agent/command/skill files (and to the kit's openspec-templates/, openspec/backlog.md, and the project's c

• stage-choreography

  The canonical main-loop procedures every QRSPI stage command runs -- run-mode establishment, precondition/approval check, commit step, next-stage handoff -- plus the hard-stop procedure, backlog atomicity, and the stage-specific gate notes. Orchestrator-only; load it in a /qrspi:* command body, neve

• vertical-slice

  How to decompose a change into vertical slices (mock service → UI → real DB) rather than horizontal layers. Load this during the Slices (V) stage when you are about to write the slice plan.

• workflow

  The eight-stage QRSPI workflow (Questions, Research, Design, Structure, Slices, Plan, Implement, PR) used on top of OpenSpec -- the stage model, the per-stage Read Matrix, and the divergence rubric. Load this when you need to know what stage you are in, what the next stage is, or why a stage exists.