Stats
Actions
Tags
From lgtm
Generate a narrated walkthrough for an LGTM review session. Produces an ordered sequence of logical-change stops with titles, narratives, and artifact references. Use when the user asks to build a walkthrough, or when /lgtm prepare chains into this skill.
How this skill is triggered — by the user, by Claude, or both
Slash command
/lgtm:walkthroughThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Generate a narrated walkthrough for an active LGTM review session. Calls the
Generate a narrated walkthrough for an active LGTM review session. Calls the
walkthrough-author agent which writes a markdown file, then set_walkthrough
parses and submits it.
None. set_walkthrough auto-registers the project if needed. If you want the
review UI claimed for notifications, call claim_reviews separately (via the
lgtm skill). Walkthrough is independent of analysis — it can run with or
without /lgtm analyze having been run first.
gh pr view --json baseRefName -q .baseRefName. If it succeeds,
fetch it with git fetch origin <branch> and use origin/<branch> as base.main (or master if main doesn't exist).Spawn the walkthrough-author agent. Pass:
REPO_PATH: <repo path>
BASE_BRANCH: <base branch>
OUTPUT_PATH: /tmp/lgtm-walkthrough.md
Call set_walkthrough with:
repoPath: the repo pathwalkthroughPath: /tmp/lgtm-walkthrough.mdIf the tool returns an error (parse failure, validation error), read the file to diagnose, ask the agent to fix, and retry.
On success, tell the user how many stops were generated and that the walkthrough
is available in the review UI (press W).
npx claudepluginhub tomplex/lgtm --plugin lgtmCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.