Skill Claude CodeCodex
Calibrated interrogation of a fuzzy idea before it becomes a spec. Asks one question at a time, recommends an answer, and lands each answer in §G (goal) or §C (constraints) — unknowns parked as ? items, never guessed. The cheapest place to kill a bad idea is before §T exists. Triggers when the user has a vague idea…
Skill Claude CodeCodex
Gather external knowledge the spec needs and distill it into §R — the durable research log — so build grounds in facts instead of hallucinating library behavior. Each finding cites a source; unsourced claims are flagged, never written as fact. Triggers when a spec decision hinges on a library/API/best practice the…
Skill Claude CodeCodex
Adversarial senior review of the spec before any code is written. Constructs a skeptical reviewer whose authority comes from the codebase, §R research, and live best-practice — then tries to REFUTE the spec, not rubber-stamp it. Every finding cites evidence (file:line or source); unverifiable ones are flagged.…
Skill Claude CodeCodex
Create, amend, or backprop bugs into SPEC.md at repo root. Sole mutator of the project spec. Triggers when the user asks to write a spec, start a new spec, distill a spec from existing code, add invariants, amend sections (§G, §C, §I, §V, §T, §B), or record a bug via backprop. Common phrasings: "write the spec…
Instructions file CodexOpenCode
Instructions for JuliusBrussee/caveman-code, covering development rules, first message, code quality, commands and github issues.
Instructions file
Instructions for JuliusBrussee/caveman-code, covering cave cli — caveman-cli, packages, key commands, context hierarchy and conventions.
Agent
Adversarial review of a plan or implementation. Pokes holes; assumes the author was over-confident.
Agent
Read-only codebase reconnaissance. Returns a compressed inventory another agent can act on without re-reading the tree.
Agent
Execute a concrete plan or patch description by editing files in an isolated git worktree.
Agent
Critique a diff or a slice of code. Returns prioritized findings with file:line citations and suggested fixes.
Agent
Run the project's test suite, parse failures, and report root cause + suggested fix per failing test.
Command
Remove obvious dead code, unused imports, and stale TODOs.
Command
Stage and commit the current working tree with a conventional-commit message.
Command
Explain a file, function, or concept in the codebase.
Command
Run the project's type-checker and fix the smallest set of errors.
Command
Start an autonomous goal loop in the background. Usage. /goal.
Command
Summarise recent git history.
Command
Walk through an opinionated multi-step code migration.
Command
Profile or analyse a hot path and propose a measured speedup.
Command
Review the diff on the current branch against main (or a target branch).
Command
Security review of the current diff (or the whole repo, if --all).
Command
Run the project's test suite and triage failures.
Skill Claude CodeCodex
Example skill loaded from resourcesdiscover.
Agent
Creates implementation plans from context and requirements.