agentsclimarketplace

Code tour

Skill tmj-90/gaffer/runner/skills/code-tour

Self-hosted AI coding factory — sandboxed agents deliver tickets to merged code, gated by a human in a dashboard. Local-first, cost-transparent, human-in-the-loop.

Install
npx -y skills add tmj-90/gaffer --skill code-tour

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

  • 0 stars0 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.

What its author says it does

Copied from the file, not written here

Use when asked to create a code walkthrough, onboarding tour, architecture tour, PR review tour, or any structured explanation of how a codebase works. Triggers on "create a code tour", "onboarding tour", "how does X work", "explain the codebase", "architecture walkthrough", "PR tour", or "contributor guide". Outputs a CodeTour `.tour` JSON file for the VS Code CodeTour extension, with a plain-markdown fallback for environments without it.

SKILL.md

5.1 KB, as published. Nobody here has run it

Create persona-targeted, file-anchored code tours

A great tour is a narrative — a story told to a specific person about what matters, why it matters, and what to do next. Every file path and line number must be real and verified.

Persona selection

Infer silently from the request:

User saysPersonaDepth
"tour for this PR"pr-reviewerstandard
"why did X break" / "RCA"rca-investigatorstandard
"onboarding" / "new joiner"new-joinerstandard
"quick tour" / "vibe check"vibecoderquick
"architecture"architectdeep
"security" / "auth review"security-reviewerstandard
(no qualifier)new-joinerstandard

Depth guidelines:

  • Quick: 5–8 steps, high-level, 1–2 sentences per step.
  • Standard: 10–20 steps, full narrative, 3–5 sentences per step.
  • Deep: 20–40 steps, links between steps, covers edge cases and design rationale.

Tour file format

{
  "$schema": "https://aka.ms/codetour-schema",
  "title": "Tour title — persona",
  "description": "One sentence: what this tour covers and who it's for.",
  "steps": [
    {
      "file": "src/index.ts",
      "line": 1,
      "title": "Entry point",
      "description": "The application starts here. The `main()` function wires together the three primary subsystems..."
    }
  ]
}

Tours live in .tours/<name>.tour in the repo root.

Markdown fallback. The .tour JSON is the primary output. When the CodeTour extension or VS Code isn't available (CI, a headless agent, a reviewer reading on GitHub), also emit — or fall back to — a plain-markdown version: a numbered list of path/to/file.ts:42 — one-line description of what happens here, one entry per step, in the same order as the tour. It carries the same narrative and is readable anywhere, so the tour is never blocked on a specific editor being installed.

Step writing principles

  1. Start with the entry point. Request handling / server startup / CLI entry / main module — wherever execution begins.
  2. Follow the call graph. Each step ends by saying where the tour goes next and why. No teleporting.
  3. Name what matters. Call out non-obvious decisions ("this uses a singleton because…"); don't describe what the code literally says.
  4. Speaker notes for complexity. For a step covering a subtle invariant, write it as if you're pair-programming — "Watch out for X here because Y."
  5. End with a summary step. "You've seen the full request lifecycle. The key files are A, B, and C. Start with A when you need to change X."

Steps

  1. Explore the repo. List root directory; read README; identify language(s), framework, entry points; map folder structure 1–2 levels deep. Every path in the tour must exist.
  2. Infer persona and depth. From the request; default to new-joiner + standard.
  3. Plan the narrative. 5-line outline before writing steps — where do we start, what's the arc, where do we end?
  4. Write steps. Verify each file path and line number before including. Line numbers must point to something meaningful (a function signature, a key conditional, a type definition).
  5. Write the summary step. Recap the key files and where to start for the most common change types.
  6. Output the .tour file. Valid JSON; placed in .tours/<descriptive-name>.tour. Emit the markdown fallback alongside it (or in place of it when VS Code isn't part of the workflow).
  7. Verify — no editor required. Parse the .tour JSON to confirm it's valid, then for every step check the file exists and the line is in range and points at something meaningful (grep/read the file at that line). This works headless; opening the tour in VS Code with the CodeTour extension is a nice final confirmation, not a prerequisite.

Review checklist

  • Every file path verified — no paths to files that don't exist.
  • Line numbers meaningful — pointing to a function signature or key line, not a blank line.
  • Narrative flows — each step connects to the next; no teleporting between unrelated files.
  • Persona-appropriate depth — quick tours don't rabbit-hole; deep tours don't skip the hard parts.
  • Summary step present — what to read first for the most common change type.
  • Valid JSON.tour file parses without errors.
  • Editor-independent — a markdown fallback exists and verification passed without opening VS Code.

Rules

  • A tour with a wrong file path is worse than no tour — verify every path before committing.
  • Only create .tour files — never modify source code to accommodate a tour.
  • If the repo has fewer than 5 source files, create a quick-depth tour regardless of persona.

Keep looking

Skills are one crate of 328,083. 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.