spec-browser
Use when there are many tech specs under specs/ and you need to navigate them as one doc site, generate a catalog agents can reference, or answer "what specs do we have / where is the spec for X". Triggers include "browse specs", "spec index", "list all specs", "spec site", "navigate specs".
- License 许可
- MIT
- 最近更新
- 2026-11-03
- Stars
- ★ 2
- Forks
- 0
01 前言 Front Matter
- name
- spec-browser
- effort
- low
- description
- Use when there are many tech specs under specs/ and you need to navigate them as one doc site, generate a catalog agents can reference, or answer "what specs do we have / where is the spec for X". Triggers include "browse specs", "spec index", "list all specs", "spec site", "navigate specs".
- when_to_use
- Reach for this once a repo accumulates several specs and a single spec folder view (technical-spec's own preview) is no longer enough — you want the whole specs/ tree at once, plus a plain-markdown catalog for agents that have no search plugin. Not for an approve/revise decision on one plan; use plan-review.
- argument-hint
- [specs-root]
02 资源文件 2
- 脚本references/build-index.js5.2 KB
- 参考文档references/package.json25 B
03 SKILL.md 正文
/spec-browser
Turn a growing specs/ tree into (1) a plain-markdown catalog any agent
can read or grep, and (2) a browsable doc site across every spec, reusing
plan-review's renderer and folder-grouped navigation. No external services,
no plugins, no build step.
Two outputs, two audiences
| Output | Audience | Form |
|---|---|---|
specs/index.md |
agents + humans | one markdown table: spec title, purpose, path |
| plan-review viewer | humans | live doc site over the whole specs/ tree |
The catalog is the portable, dependency-free way for an agent to discover and
reference specs — read or grep one file. It is deliberately not tied to any
search MCP, so it works in plain Claude Code, Codex, or any compatible agent.
Steps
1. Build / refresh the catalog
Locate this skill's generator and run it against the specs root (default
specs):
node <path-to>/spec-browser/references/build-index.js [specs-root]
It walks the tree, treats every directory containing an index.md as a spec,
extracts the title (# heading) and purpose (first > note or first line),
and writes <specs-root>/index.md. Re-run it whenever specs are added or
renamed — it overwrites the catalog in place.
2. Browse the tree (optional)
To read/annotate the whole set in the browser, launch the plan-review server
in directory mode against the specs root — it now recurses the tree and groups
the Contents nav by spec folder, with index.md (the catalog) as the landing
page:
This step needs the condux plugin's plan-review skill installed — the server ships with it, not with spec-browser. Step 1 (the catalog) works without it.
# find the server shipped with the plan-review skill
find ~/.claude ~/.codex ~/.agents -name annotate-server.js -path '*plan-review*' 2>/dev/null | head -1
node /path/to/plan-review/references/annotate-server.js specs
If the find prints nothing, plan-review isn't installed — say so and offer
the catalog (specs/index.md) instead of silently skipping the browse step.
Same annotation UI as plan/spec review: notes land in
specs/review.feedback.md. This is browsing, not a decision gate — no
approve/deny polling needed unless you want it.
Notes
- The catalog is a generated artifact — the header marks it "do not edit by hand." Editing a spec means editing that spec's own files, then re-running step 1.
- Pairs with
technical-spec(which writes individual specs) — this skill is the index over all of them. For a single spec, use plan-review's directory mode (technical-spec's old standalone preview is retired); use spec-browser when you have many. - Keep
specs/index.mdcommitted so agents in a fresh checkout can reference the catalog without running anything.
04 同仓库更多 Skills
orphan
暂无简介
no-blurb
暂无简介
workflow
Routes any dev task into the right execution tier (Small/Medium/Large) and carries condux's operating rules — proportional effort, lazy skill loading, soft gates, implement-yourself-by-default.
usage-trace
Checks and manages durable, local usage capture — ingest, health status, and the opt-in limit snapshot — for Claude Code and Codex. Capture itself runs automatically at session start.
toolkit-skill-standards
Use when writing or reviewing a SKILL.md for jabworks/agentic-toolkit — frontmatter budgets, the trigger contract ("Use when…" description or when_to_use field), progressive disclosure into references/, and keeping a new description out of existing skills' trigger space. Triggers include "review this skill description", "write the frontmatter", "does this collide with an existing skill", "where should this content live". Not for scaffold/sync mechanics; use toolkit-foundry.
toolkit-research-frontier
Use when planning improvement work on jabworks/agentic-toolkit itself — open problems (trigger-collision automation, YAML-strict frontmatter checks, docs-catalog enforcement), existing assets, the next three concrete steps, and the four-front library-health campaign with commands and success criteria. Triggers include "let's improve our toolkit", "let's improve this repo", "what should we improve next", "toolkit roadmap", "open problems", "run the health campaign".