Finishing a Development Branch
obra/superpowers
Walks the last step of a branch: confirm tests pass, detect the git environment, ask how to integrate, carry out your choice and clean up the worktree.
Review technical spec documents from completeness, feasibility, risk, and code consistency perspectives.
Install Review Spec
$ npx skills add sd0xdev/sd0x-harness --skill review-spec -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install sd0xdev/sd0x-harness review-spec --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/review-spec .claude/skills/review-spec && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "review-spec" agent skill from https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-spec into .claude/skills/review-spec/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "review-spec", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-specType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add sd0xdev/sd0x-harness --skill review-spec -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install sd0xdev/sd0x-harness review-spec --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/review-spec .agents/skills/review-spec && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "review-spec" agent skill from https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-spec into .agents/skills/review-spec/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "review-spec", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add sd0xdev/sd0x-harness --skill review-spec -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install sd0xdev/sd0x-harness review-spec --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/review-spec .cursor/skills/review-spec && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "review-spec" agent skill from https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-spec into .cursor/skills/review-spec/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "review-spec", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/sd0xdev/sd0x-harness.git --path skills/review-spec--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add sd0xdev/sd0x-harness --skill review-spec -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install sd0xdev/sd0x-harness review-spec --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/review-spec .gemini/skills/review-spec && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "review-spec" agent skill from https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-spec into .gemini/skills/review-spec/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "review-spec", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install sd0xdev/sd0x-harness review-specInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add sd0xdev/sd0x-harness --skill review-spec -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/review-spec .github/skills/review-spec && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "review-spec" agent skill from https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-spec into .github/skills/review-spec/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "review-spec", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add sd0xdev/sd0x-harness --skill review-spec -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install sd0xdev/sd0x-harness review-spec --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/sd0xdev/sd0x-harness.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/review-spec .opencode/skills/review-spec && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "review-spec" agent skill from https://github.com/sd0xdev/sd0x-harness/tree/main/skills/review-spec into .opencode/skills/review-spec/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "review-spec", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
review-spec- GitHub stars
- 192
- Token cost
- ~2.1k tokens
- SKILL.md length
- 1,039 words
- Files
- 1
- Skills in repo
- 91
- Repo updated
- First seen
- Licence
- MIT
At a glance
Review technical spec documents from completeness, feasibility, risk, and code consistency perspectives.
- Works in 2 steps: Read the full document: cat ${FILE_PATH} → If long: cat ${FILE_PATH} | head -300…
- Development work in your project
- SKILL.md covers Trigger, When NOT to Use, Relationship to… and Codex Dispatch, plus 10 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Review Spec is an agent skill from sd0xdev/sd0x-harness. Review technical spec documents from completeness, feasibility, risk, and code consistency perspectives.
Its SKILL.md is about 2.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Development. The repository describes itself as: The harness layer for Claude Code — a reference implementation of harness engineering with hook-enforced dual review, state-machine gates that survive context compaction, and… The licence is MIT.
When your agent uses it
- Development work in your project
Example prompts
- “/review-spec”
Requirements
- Pre-approved tools (allowed-tools): Read, Grep, Glob, Bash(git:*), Bash(node:*), Write
Workflow steps
2 steps, taken from the first numbered list in SKILL.md.
- Read the full document: cat ${FILE_PATH}
- If long: cat ${FILE_PATH} | head -300 then cat ${FILE_PATH} | tail -200
What it can do on your machine
Read from SKILL.md and the folder at commit c9a2036. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves these tools, so the agent can use them without asking each time:
ReadGrepGlobBash(git:*)Bash(node:*)Write
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Review Spec loads about 2.1k tokens when it runs. Until then it costs about 29 tokens; SKILL.md has 1,039 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from sd0xdev/sd0x-harness at commit c9a2036, republished under its MIT licence (© sd0xdev). 1,039 words, ~2,102 tokens.
Download SKILL.mdSave it as
.claude/skills/review-spec/SKILL.md (or your agent's skills folder).- name
- review-spec
- description
- Review technical spec documents from completeness, feasibility, risk, and code consistency perspectives.
- allowed-tools
- Read, Grep, Glob, Bash(git:*), Bash(node:*), Write
Review Spec
Trigger
- Keywords: review spec, spec review, tech spec review, review-spec
When NOT to Use
- Code review (use
/codex-review-fast) - General document review (use
/codex-review-doc) - Writing a new spec (use
/tech-spec)
Relationship to /codex-review-doc
Both are doc-plane producers of the same gate, dispatched over the same
Codex exec transport (@skills/codex-code-review/references/codex-transport.md) and emitting the same sentinel pair. They differ only in review
depth and dimensions: /review-spec is the design-landing depth (completeness, feasibility,
risk, code consistency, test strategy), /codex-review-doc is the general document depth.
Loop mechanics, severity calibration and [NIT_DEFERRED] handling are shared — see
@skills/doc-review/SKILL.md.
Why not an Agent dispatch. The gate verdict is behaviour-layer: it comes from the reviewer's
report, and review-state.js note records it (@rules/auto-loop.md § Enforcement — nothing parses
reviewer output; hooks are reminders). What makes a verdict usable is that a contract-aware
reviewer produced it against this family's template and sentinels. A built-in agent dispatched
ad hoc is not that reviewer and its output closes nothing. Dispatching Codex over the shared
transport is what makes the verdict this skill's to note. See @rules/auto-loop.md § Review Dispatch.
Codex Dispatch
Bind every placeholder before writing prompt.md. The body below is body-only, so nothing
evaluates an expression inside it: ${FILE_PATH} and ${PROJECT_ROOT} must carry real values by the
time the file is written, or the cat ${FILE_PATH} instructions reach Codex as literal text and
cannot be run.
You are a senior technical spec reviewer. Perform a Document Review of the technical specification below, at design-landing depth.
Document Info
- Path: ${FILE_PATH}
- Type: technical specification
- Project root: ${PROJECT_ROOT}
⚠️ Important: You must independently read and research the project ⚠️
Do NOT expect pre-provided file content. Read the spec and research the project yourself using your sandbox access. The spec makes concrete claims about this repository — verify them against the actual files rather than taking them on trust.
Document Reading (Priority)
- Read the full document:
cat ${FILE_PATH} - If long:
cat ${FILE_PATH} | head -300thencat ${FILE_PATH} | tail -200
Code-Documentation Consistency Research
- Project structure, discovered rather than assumed:
lsat the repository root, then the directories it actually shows — do not assume asrc/layout; many repositories, this one included, have none - Search for every file, function, flag and command the spec names:
grep -rn "keyword" . -l --include="*.ts" --include="*.js" --include="*.sh" | head -10 - Read related files:
cat <file-path> | head -100 - Verify: do referenced files exist? Are names correct? Do described behaviours match code?
Review Dimensions
| # | Dimension | Checks |
|---|---|---|
| 1 | Completeness | Are requirements, scope, risks and work breakdown all present and specific |
| 2 | Feasibility | Can this be built as described, with the dependencies it names |
| 3 | Risk Assessment | Are the real failure modes identified, and does each have a bound |
| 4 | Code Consistency | Do referenced files/functions exist and behave as described (verify with grep/cat) |
| 5 | Test Strategy | Is every acceptance criterion mapped to evidence; are guards two-directional |
Severity Calibration ⚠️
A 🔴 blocks the document and costs a full review round. Reserve it for defects that would mislead a reader into building the wrong thing:
| Mark 🔴 | Do NOT mark 🔴 |
|---|---|
| A described file, function, flag or command that does not exist | Wording that could be clearer |
| A described behaviour that contradicts what the code actually does | A section you would have structured differently |
| A security or data-handling design that is wrong or unsafe | A missing section that no rule requires |
| An internal contradiction — two passages that cannot both be true | Prose where a table would be tidier |
| A broken cross-reference, or a step whose stated dependency is not met by its own ordering | Hypothetical future concerns not present in the change |
Do not manufacture findings to fill a section. An empty 🔴 section is a normal outcome.
Show full SKILL.md (433 more words)Show less
Output Format
Your report must begin with the literal line ## Document Review. Nothing parses it — the
verdict is behaviour-layer and review-state.js records only an explicit note
(@rules/auto-loop.md § Enforcement). The header matters for the reader: it is what tells a
document review apart from a code or security review in a transcript.
Document Review
Review Summary
| Dimension | Rating (1-5⭐) | Notes |
|---|---|---|
| Completeness | ... | ... |
| Feasibility | ... | ... |
| Risk Assessment | ... | ... |
| Code Consistency | ... | ... |
| Test Strategy | ... | ... |
🔴 Must Fix (blocking — see Severity Calibration)
- [Section/Line] Issue description -> Fix recommendation
(Write None if there are none.)
🟡 Suggested Changes (non-blocking)
- [Section/Line] Issue description -> Fix recommendation
⚪ Optional Improvements
- Suggestion
Deferred Findings
For every 🟡 and ⚪ above, emit one line here, starting at column 0:
[NIT_DEFERRED] <file:line> | <issue> | reason: sub-threshold-doc | <ISO8601 UTC>Do not reorder the fields and do not use a different tag. Omit this section entirely if there are no 🟡 or ⚪ items.
Gate
End the report with exactly one verdict terminal, alone at column 0 on the final line — the
same rule as @skills/doc-review/references/review-loop-doc.md, and what
scripts/validate-family-sentinel.js doc accepts. As list items they are not a legal terminal.
- No 🔴 findings → final line
✅ Mergeable - Any 🔴 finding → final line
⛔ Needs revision
Dispatch this body per @skills/codex-code-review/references/codex-transport.md § Start; the transport pins the sandbox and approval policy, so nothing is chosen here. Save the returned threadId — loop re-review continues the same thread per that reference's § Resume; see @skills/doc-review/references/review-loop-doc.md.
Task
Document to Review
$ARGUMENTSIf no path is given, auto-detect: git-modified 2-*.md under docs/features/ → staged
.md → newest tech spec. Multiple candidates: list them and ask which to review.
Gate
| Sentinel | Meaning |
|---|---|
✅ Mergeable | No 🔴 items — the spec may proceed to implementation |
⛔ Needs revision | 🔴 items present — fix, then re-review on the same thread |
These are the doc-plane sentinels (@rules/auto-loop.md § Gate Sentinels). No hook parses them —
the verdict is behaviour-layer and review-state.js records only an explicit note. One thing does
read them mechanically, and it is not a hook: scripts/validate-family-sentinel.js doc validates a
fallback carrier's raw report before its verdict may be adopted. That is why the shape is fixed
in both directions — a paraphrase reads as no verdict to a human, and fails the validator outright
when a fallback produced it. Emit them verbatim.
🔴 only. 🟡 and ⚪ are non-blocking: log them via [NIT_DEFERRED] and proceed
(@rules/auto-loop.md § Sub-Threshold Findings). Round cap and its ## Max Rounds override come from the shared contract (@rules/auto-loop.md § Tiers); still failing
→ report the blocker rather than spending another round.
Verification
- Dispatched over the Codex exec transport (§ Start), not an Agent
- The prompt asks for a
Document Review, and the report opens with that header - Codex verified code-documentation consistency independently
- Gate is one of
✅ Mergeable/⛔ Needs revision
References
- Shared loop mechanics and severity model:
@skills/doc-review/SKILL.md - Re-review template:
@skills/doc-review/references/review-loop-doc.md - Dispatch contract:
@rules/codex-invocation.md
© sd0xdev, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in skills/review-spec of sd0xdev/sd0x-harness.
Open the folder on GitHubat commit c9a2036
Compare with similar skills
Review Spec next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Review Spec this skillsd0xdev/sd0x-harness | 192 | — | ~2.1k | Automated safety check: Pass | MIT | |
| Finishing a Development Branchobra/superpowers | 296k | 5 repos | ~1.9k | Automated safety check: Pass | MIT | |
| Typescript Advanced Typesrolling-scopes/rsschool-app | 10k | 24 repos | ~4.2k | Automated safety check: Pass | MPL-2.0 | |
| PR Babysitteropeninterpreter/openinterpreter | 69k | 3 repos | ~4.2k | Automated safety check: Pass | Apache-2.0 | |
| Code Review ChecklistshareAI-lab/learn-claude-code | 78k | 5 repos | ~1.1k | Automated safety check: Pass | MIT | |
| Greplooponyx-dot-app/onyx | 32k | 4 repos | ~3.3k | Automated safety check: Pass | MIT |
Similar skills
Typescript Advanced Types
rolling-scopes/rsschool-app
Master TypeScript's advanced type system including generics, conditional types, mapped types, template literals, and utility types for building type-safe applications.
PR Babysitter
openinterpreter/openinterpreter
Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.
Code Review Checklist
shareAI-lab/learn-claude-code
Reviews code against a five-part checklist covering security, correctness, performance, maintainability and testing, and reports findings in a fixed format.
Greploop
onyx-dot-app/onyx
Iteratively improves a PR (GitHub), MR (GitLab), or shelved changelist (Perforce) until Greptile gives it a 5/5 confidence score with zero unresolved comments.
Guidelines
akash-network/node
Behavioral guidelines to reduce common LLM coding mistakes. An agent skill from akash-network/node.
More from sd0xdev/sd0x-harness
All 91 skills in this repoAdr
sd0xdev/sd0x-harness
Write an Architecture Decision Record (ADR) for a feature — Context / Decision / Status / Consequences / Alternatives, filed as docs/features/<feature/adr-<NNN-<title.md with a 3-digit zero-padded…
Load PR Review
sd0xdev/sd0x-harness
Load GitHub PR review comments into AI session — analyze, triage, plan.
Next Step
sd0xdev/sd0x-harness
Change-aware next step advisor. An agent skill from sd0xdev/sd0x-harness.
Obsidian CLI
sd0xdev/sd0x-harness
Obsidian vault integration via official CLI. An agent skill from sd0xdev/sd0x-harness.
Orchestrate
sd0xdev/sd0x-harness
Agent-driven workflow orchestration (v1 report-only). An agent skill from sd0xdev/sd0x-harness.
PR Comment
sd0xdev/sd0x-harness
Post friendly review comments to a GitHub PR — prepare locally, preview, then submit as atomic review.
Related
Categories
Questions about Review Spec
What does Review Spec do?
Review technical spec documents from completeness, feasibility, risk, and code consistency perspectives. Review Spec is an agent skill from sd0xdev/sd0x-harness. Review technical spec documents from completeness, feasibility, risk, and code consistency perspectives.
When should I use Review Spec?
Review Spec fits situations like: development work in your project.
How do I install Review Spec in Claude Code?
Run `npx skills add sd0xdev/sd0x-harness --skill review-spec -a claude-code`. Or copy the skill folder (skills/review-spec in sd0xdev/sd0x-harness) into .claude/skills/review-spec in your project. Claude Code loads it when a task matches its description.
How do I install Review Spec in Codex?
Run `npx skills add sd0xdev/sd0x-harness --skill review-spec -a codex`. Or copy the skill folder (skills/review-spec in sd0xdev/sd0x-harness) into .agents/skills/review-spec in your project. Codex loads it when a task matches its description.
Can I use Review Spec in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add sd0xdev/sd0x-harness --skill review-spec -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/review-spec, .gemini/skills/review-spec, .github/skills/review-spec and .opencode/skills/review-spec in your project.
What does Review Spec need to run?
SKILL.md names no scripts, command-line tools or credentials: Review Spec is instructions for the agent only. Its frontmatter pre-approves these tools: Read, Grep, Glob, Bash(git:*), Bash(node:*), Write.
Does Review Spec access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Review Spec safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Review Spec use?
Review Spec is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Review Spec use?
About 2.1k tokens (SKILL.md is roughly 8.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Review Spec?
Skills that share tags, products or a category with Review Spec: Finishing a Development Branch (obra/superpowers, 296k stars), Typescript Advanced Types (rolling-scopes/rsschool-app, 10k stars), PR Babysitter (openinterpreter/openinterpreter, 69k stars) and Code Review Checklist (shareAI-lab/learn-claude-code, 78k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Review Spec?
sd0xdev (a GitHub user) maintains it in sd0xdev/sd0x-harness, which has 192 GitHub stars. The repository holds 91 skills in this directory. The repository was last updated on October 6, 2026.
Source: sd0xdev/sd0x-harness on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.