write-review
Review a SpecForge spec as a reviewer agent, and file the review as a child spec of the spec being reviewed.
Skill by NitinJ
Chalk Labs
Need help deploying write-review?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/nitinj-write-review
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing, productivity, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/write-review
Tools it pre-approves 6
From allowed-tools in SKILL.md: the agent may use these without asking you each time.
Read,Write,Edit,Bash,Glob,Grep
What's inside
- Skill name
write-review- Description
- Review a SpecForge spec as a reviewer agent, and file the review as a child spec of the spec being reviewed. Use when asked to "review spec <id>", "critique this spec", "give a second opinion on <spec>", or when a workflow hands you a spec to review. Creates the review with `specforge review`, writes a verdict and severity-ranked findings, credits you as a reviewer of the reviewed spec, and passes the verify gate. Not for answering comments on a spec you own: that is review-spec.
- License
- MIT
- Allowed tools
- Read,, Write,, Edit,, Bash,, Glob,, Grep
- Instructions
- 78 lines in SKILL.md · view on GitHub
Before installing
- Pre-approves some shell commands
Files (1)
- SKILL.md
SKILL.md (first lines)
# write-review
A review is a spec of its own, filed under the spec it reviews. It gets its own
URL, status and comment threads, and the reviewed spec's **Child specs** drawer
lists every review it has had. The reviewed spec shows you as a reviewer on the
home page and in its header.
`${CLAUDE_PLUGIN_ROOT}` below denotes the installed plugin directory. Claude and
Pi export it; Codex provides the exact value in SpecForge SessionStart context.
Substitute that value in every path and shell command.
## The cadence
- **One review is one child spec.** Reviewing again after the author changed
the spec is a new review, never an edit of the old one, so the history stays.
- **Never edit the reviewed spec**, and do not leave comments on it. The review
is the channel. The author decides what changes.
- **Say so if you are its author.** The home page and header show who wrote it.
A self-review is allowed when asked, and its scope section says it is one.
## 1. Read the spec you are reviewing
Resolve what you were given to a spec id first. A store id is 10 hex
characters; anything else, look up with:
```
node "${CLAUDE_PLUGIN_ROOT}/lib/specforge-cli.mjs" listall
```
Then read it by map and section rather than as one file:
```
node "${CLAUDE_PLUGIN_ROOT}/lib/spec-nav-cli.mjs" map --spec ~/.specforge/specs/<id>/spec.html
node "${CLAUDE_PLUGIN_ROOT}/lib/spec-nav-cli.mjs" section <sectionId> --spec ~/.specforge/specs/<id>/spec.html
```
Check its claims against whatever it rests on: the code, the sibling specs it
links, the request it answers. A review that only read the spec says so.
## 2. Create the review
```
node "${CLAUDE_PLUGIN_ROOT}/lib/specforge-cli.mjs" review <id> --model <your model id>
```
`--model` is your own model id, exactly as your harness names it (for example
`claude-opus-5`, `gpt-6-astra`, `glm-5.3-flash`). The harness is detected. The
command creates the child spec, credits you as its author, and credits you as a
reviewer of `<id>`.
It prints what `create` prints: `{ id, htmlPath, url, skeleton, prompts,
language, author, reviewOf, reviewers }`. Author into `htmlPath`, section by
section, as the `create-spec` skill describes. **`language` is the writing
contract in force; follow it.** **Read `prompts` before writing each section.**
## 3. Write the four sections
Verdict, what you reviewed, findings, questions for the author. The prompts
carry the rules. Keep the whole review short: five blocking and major findings
are worth more than twenty minor ones. Link each finding to the section it is
about, as `/spec/<reviewed id>#<section id>`.
## 4. Pass the gate
```
node "${CLAUDE_PLUGIN_ROOT}/lib/specforge-cli.mjs" verify <reviewId> --json
```
The same loop as `create-spec` step 4: fix what fails, judge the ask rules,
three rounds at most.
## 5. Hand off
Print the review's URL and its verdict in one line. The review is attached to
your session, so comments on it reach you; arm delivery the way `create-spec`
step 5 describes.From NitinJ/specforge, MIT.
More from NitinJ/specforge
- 0.579score
generate-template
NitinJ
Write a spec template from a prompt.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
export
NitinJ
Export a spec from the store to Google Docs.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
list-specs
NitinJ
List SpecForge specs — either every spec in the store (mode "all") or just the specs attached to this session (mode "mine").
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
migrate-spec
NitinJ
Move one existing spec onto the SpecForge component library.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
tune-templates
NitinJ
Optimize the SpecForge spec templates from the review-comment history.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
export-md
NitinJ
Export a spec from the store as a GitHub-flavoured markdown file, plus a sidecar directory holding its diagrams.
SkillBFreshness: B★ 5Claude Code +9 - 0.575score
create-spec
NitinJ
Author a new house-style spec into the SpecForge store.
SkillBFreshness: B★ 5Claude Code +9 - 0.575score
convert-spec
NitinJ
Convert an existing design/spec file into a SpecForge spec in the store.
SkillBFreshness: B★ 5Claude Code +9 - 0.575score
review-spec
NitinJ
Process a submitted batch of human comments on a spec in the store: reply to each thread inline and amend the spec accordingly.
SkillBFreshness: B★ 5Claude Code +9 - 0.574score
start-specforge
NitinJ
Start or reuse the local SpecForge daemon and return its browser index URL.
SkillBFreshness: B★ 5Claude Code +9
Signals
- Stars
- 5
- Forks
- 0
- Open issues
- 1
- Last commit (this folder)
- Sep 18, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/NitinJ/specforge
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from NitinJ
- 0.579score
generate-template
NitinJ
Write a spec template from a prompt.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
export
NitinJ
Export a spec from the store to Google Docs.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
list-specs
NitinJ
List SpecForge specs — either every spec in the store (mode "all") or just the specs attached to this session (mode "mine").
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
migrate-spec
NitinJ
Move one existing spec onto the SpecForge component library.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
tune-templates
NitinJ
Optimize the SpecForge spec templates from the review-comment history.
SkillBFreshness: B★ 5Claude Code +9 - 0.579score
export-md
NitinJ
Export a spec from the store as a GitHub-flavoured markdown file, plus a sidecar directory holding its diagrams.
SkillBFreshness: B★ 5Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get write-review running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
write-review skill. Agent Mart. https://agentmart.dev/skills/nitinj-write-review (last updated 2026-10-24)