lore

Drive a feature end-to-end across multiple protocol sessions — the coordinator role's protocol home

Create a GitHub pull request from the current branch, deriving the PR body from the associated work item's plan and notes.

Holistic multi-lens PR review with adaptive lens selection, cross-lens synthesis, and structured findings; --self runs the same pipeline as an author's self-review. Use individual lens skills (/pr-correctness, /pr-security, etc.) for focused single-concern analysis.

Create a technical specification — `/spec short` for single-pass plans, `/spec` for full team-based investigation

Check project status, remaining tasks, and session context — USE FIRST when asked 'what's left', 'what should I do', 'remaining work', or status questions. Also: create, update, archive, search work items.

Focused lens review: trace the impact of PR changes on code outside the diff. Use /pr-review for integrated multi-lens coverage.

Focused lens review: trace logic paths for correctness bugs in a PR. Use /pr-review for integrated multi-lens coverage.

Focused lens review: evaluate interface clarity, naming, and abstraction coherence in a PR. Use /pr-review for integrated multi-lens coverage.

Focused lens review: detect capability loss from deletions and modifications in a PR. Use /pr-review for integrated multi-lens coverage.

Focused lens review: evaluate security vulnerabilities, edge cases, and adversarial paths in a PR. Use /pr-review for integrated multi-lens coverage.

Focused lens review: evaluate test coverage, quality, and assertion rigor in a PR. Use /pr-review for integrated multi-lens coverage.

Focused lens review: evaluate thematic coherence and scope of a PR. Use /pr-review for integrated multi-lens coverage.

Focused lens review: evaluate user and developer impact of design decisions in a PR. Use /pr-review for integrated multi-lens coverage.

Review and apply accumulated protocol evolution suggestions from retro

Author-calibrated self-review: parallel lens pre-scan (Blast Radius, Security, Test Quality, Correctness, Regressions, Interface Clarity, User Impact) then a materiality gate, persisting findings into a followup sidecar for TUI triage

Structured discussion of followup findings with source-aware routing and action affordances

Interactive pair-review with knowledge-enriched turn protocol

Read all PR review feedback and create a work item with categorized findings and verification directives

Evaluate the lore memory system — run structured tests across 9 dimensions, produce scored comparable results, and log protocol evolution suggestions

Evaluate knowledge-system effectiveness after a work cycle from a source-manifested evidence pack, then atomically persist explicit diagnostic judgments and any evolution suggestions

Manage the per-project knowledge store — USE FIRST when searching for past decisions, patterns, conventions, or architecture before Grep or Explore agents. Commands: add, search, view, curate, heal, init

Capture insights to knowledge store and update conversational threads — invoke anytime to ensure nothing is lost

Bootstrap a knowledge store by exploring codebase architecture — use when starting with a new or empty project, seeding knowledge, or running /bootstrap

Execute a spec's plan with a knowledge-aware agent team — spawns workers, tracks progress, captures architectural findings

Full knowledge store renormalization — prunes stale entries, merges redundancies, rebalances category structure via orchestrated multi-agent flow