ask
CommandAsk a codebase question and get a cited, synthesised answer with a confidence rating (costs an LLM call).
Work with architectural decisions — list, inspect health, add, or confirm auto-proposed decisions.
> /plugin marketplace add repowise-dev/repowise> /plugin install repowise@repowise
description: Work with architectural decisions; list, check health, record, pin, or put background records in force. allowed-tools: Bash(repowise decision:*), Bash(repowise config list:*), Bash(repowise config explain:*), Read
Repowise keeps architectural decisions (the *why* behind the code) next to the files they govern and tracks them for staleness. A decision an agent records comes into force when its commit merges; records mined from history are background until someone puts them in force. This command drives the `repowise decision` group. `repowise config list --area decisions` prints the repository's mode (`decisions.mode`: `full`, `read_only` or `off`) and every switch, with where each value came from.
1. If `.repowise/` doesn't exist: "This repo isn't indexed yet. Run `/repowise:init` first." Stop. 2. Map `$ARGUMENTS` to a subcommand, run it, and present the result clearly.
ungoverned hotspots (high-churn files with no recorded decision). Good first call.
spans, status, and the supersession lineage.
`repowise decision add --agent <slug> --decision D --rationale R --affects <path> --evidence-commit <sha>` (e.g. `claude_code`, plus `--session <id>` if known), so the record is the agent's, not a person's. `--evidence-commit` ties it to the commit the choice was made in; an agent's record is accepted only with evidence, from this or `--evidence` on confirm. With both `--title` and `--decision` it records without prompting and prints the id (`--format json` to parse it back). Also optional: `--context`, `--alternative`, `--consequence`, `--tag`. `--alternative`, `--consequence`, `--affects`, `--tag` and `--evidence-commit` repeat. An agent's record that cites its commit comes into force when that commit merges to the default branch. **Always state a reason** — pass `--rationale` (or `--context`); a record whose body only restates its own title cannot be accepted. `--kind agreement` records a working agreement: a rule about how the work is conducted, which names no file and is not checked against the code. Bare `repowise decision add` is a guided interactive capture (~90s) for a person at a terminal.
records on the busiest production files (at most 20) to check against the code, put in force signed as you with evidence, or dismiss. Follow it as printed. With no background records it has you record the decisions behind the busiest files instead. `--agreements` prints the prompt that moves stored working agreements into CLAUDE.md, for a team that wants them there.
records not in force, each with its id and the `blockers` that stop it being accepted.
puts records in force, signed as the agent (plus `--session <id>` if known); the acceptance is recorded as the agent's, not a person's. `--decision`, `--reason`, `--scope <path>` and `--evidence <commit|file|link>` fill the blockers. The unsigned form is a person's. A repository with `decisions.agent_acceptance` set to false refuses it; then leave acceptance to a person. `confirm` refuses a decision that is already accepted; renew that one with `reaffirm`.
`needs_review` decision that still holds, renew it so it counts as current from now. Takes `--reason`, `--evidence`, `--agent <slug>` and `--preview`. A background record is refused; put it in force with `confirm`.
background record into a decision in force instead of accepting it twice. The target must already be in force: confirm it first, or merge the other way. `--preview` shows the result without writing.
(optionally `--superseded-by <id>`, `--preview`) or `repowise decision dismiss <id> --reason ...`. An agent passes `--agent <slug>` to every review command, and the signer is recorded for background records too. A pinned decision refuses both from an agent; a person passes `--force`.
law (listed by name in CLAUDE.md; agents ask a person before retiring it); `repowise decision unpin <id>` releases it. Run them when the user asks, with `--agent <slug>`.
what came into force in the last seven days and who signed, and what capture did.
user asks to turn decisions off or make them read-only; it resets the finer decision switches too, so relay what moved and the undo commands it prints. `repowise config explain <key>` says what any other decision switch does.
MCP tool (the `architectural-decisions` skill) — it returns lineage and an alignment score. This command is for *managing* the decision records themselves.
subcommands take ids and flags and never prompt, except `dismiss` (pass `--yes`).
Codebase intelligence for AI and humans: code health scores, auto-generated docs, git analytics, dead code detection, and architectural decisions via MCP.
Ask a codebase question and get a cited, synthesised answer with a confidence rating (costs an LLM call).
Triage card for files, modules, or symbols — layer, hotspot, fix history, freshness (relationships, not source bytes).
Ingest or inspect test-coverage reports (LCOV, Cobertura, Clover, Go cover profiles, JaCoCo, Salesforce Apex JSON, or repowise…
Report unreachable files, unused exports, and zombie packages, tiered by confidence.
Diagnose the Repowise setup — install, API keys, index/store drift — and optionally repair it.
Export the wiki (or architecture model) to markdown, HTML, JSON, or Structurizr DSL.
Implement tasks incrementally — build, test, verify, commit. Add "auto" to run the whole plan in one approved pass.