galdr

Use once a spec is approved and ready to become an executable task list — turns spec requirements into a wave-based plan of write-scoped, testable tasks that waves can dispatch. Invoke via route or explicit command.

Use when a plan already exists and is ready to run — dispatches each wave's tasks, reviews every return before trusting it, and gates advancement on evidence. Invoke via route or explicit command.

Use when writing a new galdr skill, editing an existing one, or running the pack's quarterly prune review.

Continuation protocol for resuming work after a break, handing off to another session, a subagent, or Codex, and parking fast under a usage limit. Invoke via route or explicit command.

Use once per repo to create or refresh docs/agents/galdr.md — the config verify, waves, review, branches, and route read for gate commands, invariants, models, and smoke details. Slash-command only, not router- or model-invoked.

Use when the user runs /galdr:core to install, uninstall, or check galdr's core rules block on this machine for Claude, Codex, or Antigravity. Run install once per agent to write the marked block into that agent's rules file (or the Claude enabled flag); uninstall to remove it; status to report state without writing.

Use when turning a fuzzy idea, a position to stress-test, or a set of already-clear requirements into a written spec — grill mode tests a position, explore mode is open-ended, synthesis mode drafts straight from clear requirements. Invoke via route or explicit command.

Use when starting work on a new branch or worktree, or when a branch's implementation is done and it needs gates, review, a smoke sheet, and a decision about merge/push. Invoke via route or explicit command.

Use when the user runs /galdr:usage-bridge to install, remove, or check galdr's statusline rate-limits bridge on this machine — the wrapper that makes the usage % and quota-park work. Run install once per machine; uninstall to fully revert; status to check.

Use before claiming any task, gate, or bugfix is done — proves the claim with a command run this message instead of trusting memory or a report. Invoke via route or explicit command.

Use when the user runs /galdr:backlog, or when a skill explicitly defers something to a later cycle — reviews, records, or resolves deferred work in docs/backlog.md, the pack's backlog file.

Use when diagnosing a bug, flaky test, unexpected behavior, or production incident, before proposing or writing any fix. Invoke via route or explicit command.

Use when a design question can be answered faster by building working throwaway code than by writing a spec or arguing about it — comparing two implementation approaches, checking how a library or API actually behaves, or testing whether an idea works before committing to a plan. User-invoked only.

Use when architecture pain is blocking current work, or for a periodic pass over a known pain area — maps the area read-only, ranks findings, stress-tests the top one, and exits into a plan. Not a standing background activity. User-invoked only.

Use when reviewing a branch, wave, or PR's changes against a SHA range — checks the diff against this repo's conventions and against the originating plan/spec, as two separate axes that are never merged into one verdict. Invoke via route or explicit command.

Use before acting on any request that will change a repo, or that asks for a design decision, debugging help, or planning — decides which skill (or no skill) handles it next.

Use before writing or changing any production code — features, bug fixes, refactors — and whenever the router announces the tdd-fast-path. Invoke via route or explicit command.

Use when unsure which galdr skill to invoke, or when two skill names seem to overlap (shape vs prototype, review vs verify, debug vs tdd, plan vs waves).