agentsclimarketplace

Calm guide

Skill Eilodon/CALM/.claude/skills/calm-guide

Use before searching, reading, or editing files in this repo (or any repo with a "calm" MCP server), or whenever unsure which CALM tool fits — a self-triggerable orientation to the CALM Stage 1-8 tool workflow (repo_overview → locate → source → callers → edit_context → edit_symbol/edit_lines → diff_impact → session_context). Also invoke manually as /calm-guide for a mid-session refresher, or whenever AGENTS.md's SessionStart auto-injection didn't happen this turn (some MCP clients — and non-Claude-Code sessions entirely — never get it).From its SKILL.md

Install
npx -y skills add Eilodon/CALM --skill calm-guide

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 15 stars15 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.0 KB, as published. Nobody here has run it

CALM MCP workflow — quick trigger

Read AGENTS.md at the project root in full now. It is the single source of truth for this workflow — this skill exists to make sure you actually reach for it at the right moment, not to hold a second copy of its content that would silently drift out of sync with it over time. If AGENTS.md doesn't exist at the project root, this isn't a CALM-instrumented repo (or you're in the wrong directory) — stop here, this skill doesn't apply.

If you only have room for the short version right now

  1. Orient — repo_overview() always first, then hotspots()/ fitness_report() as needed.
  2. Locate — locate(query) (search + file_overview + symbol_info in one call) or search(query, kind=...).
  3. Inspect — source(symbol) for a symbol-precise read, or understand(symbol) for locate+source+callers together.
  4. Trace — callers(symbol), callees(symbol), path(from, to), dependencies(path).
  5. Pre-Edit — edit_context(symbol): mandatory before touching any real code symbol, no exceptions. (A Markdown/text heading is the one exception — it never carries a blast radius to check.)
  6. Edit — edit_symbol(...)/edit_lines(...): CALM's own write path, hash-verified, risk-gated, reindexes immediately. Fall back to native Edit/Write only for a brand-new file or a path CALM doesn't index (dotdirs, target/, node_modules/, dist/, build/, etc.).
  7. Verify — diff_impact(staged=true): mandatory before any commit or push, no exceptions.
  8. Recover — session_context() after 10+ calls without progress; remember(topic, content)/recall(topic) for durable cross-session notes.

Steps 5 and 7 are the only hard-enforced ones under Claude Code (a hook denies the native equivalent without them) — the rest is strongly recommended, not blocking. Every signal, edge case, and the preset/toolset reference lives in AGENTS.md — go read it in full if you haven't this session.

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.