agentsclimarketplace
Sign in

Specdd review

Skill specdd/plugin-antigravity/skills/specdd-review

Use when Antigravity needs to review changes against active SpecDD specs.From its SKILL.md

Install
npx -y skills add specdd/plugin-antigravity --skill specdd-review

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 file declares

Copied from the file, not written here

The file declares its own license as Apache-2.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.5 KB, 714 tokens by cl100k_base, as published. Nobody here has run it

SpecDD Review

Use this skill for code-review style review under SpecDD. Do not edit files unless the user explicitly asks for changes.

Skill Scope

Use this skill only for projects that already have a root .specdd/bootstrap.md and are therefore SpecDD projects.

If .specdd/bootstrap.md is missing, do not continue with this skill. Use specdd-adopt only when the operator explicitly asks to add or adopt SpecDD in a new or existing non-SpecDD project.

Bootstrap Resolution

Ensure you know the active bootstrap chain for the target project. If it is not already reliable in the current context, read bootstrap files in this exact order:

  1. .specdd/bootstrap.md
  2. .specdd/bootstrap.project.md, if it exists
  3. .specdd/bootstrap.local.md, if it exists

Apply them in that order. Later files may narrow or specialize earlier rules, but must not silently weaken project contracts, inherited constraints, or write authority.

Do not reread bootstrap files merely because another SpecDD skill or workflow phase starts. Reread the specific bootstrap file only when the active project root changes, a bootstrap file may have changed, or you need exact bootstrap text for a decision, quote, comparison, authorization, or report.

Spec Chain Resolution

Before planning, editing, reviewing, testing, or reporting on a target, ensure the effective spec chain for that target is known:

  1. Identify the target path, task, behavior, or changed files.
  2. Resolve the effective spec chain using the active bootstrap rules.
  3. Include same-directory basename specs when applicable.
  4. Walk ancestor specs from the selected content root to the target.
  5. Read explicit References only when they affect the requested work.
  6. Read or reread the specific governing specs needed for the current decision when exact contract text is not already reliable in context.
  7. Identify the nearest local spec that grants write authority before any edit.

Do not infer authority from similar filenames, nearby files, symbols, module names, or conventions unless the active bootstrap or a governing spec explicitly defines that mapping.

Do not reread an entire chain just because a workflow phase changed. Reopen the narrowest relevant spec when you need to apply, quote, compare, authorize, report, or edit against exact text, when the target scope changes, when a governing spec may have changed, or when earlier context came only from CLI output or a summary.

Workflow

  1. Compare the diff against applicable Must, Must not, Forbids, Scenario, Tasks, and Done when entries.
  2. When useful and available, consider consulting the specdd-cli skill for CLI-assisted governing-spec discovery before judging a diff; read or reread the relevant contracts directly when exact text is needed for a finding.
  3. Check whether verification matches the behavioral risk.
  4. Report concrete findings, questions, or residual risk.

Review Output

Lead with findings, ordered by severity. For each finding, include:

  • File and line when available.
  • The violated or risky SpecDD contract.
  • The concrete behavior or regression risk.
  • The smallest likely fix.

Then include open questions or assumptions. Keep summaries secondary and brief. If there are no findings, say so clearly and mention any residual test or spec coverage gap.

Do not invent requirements beyond the active specs. If behavior is underspecified, call that out as an uncertainty rather than a violation.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 1 of the 12 instructions most review quality skills give in 714 tokens

Counted across 1,048 of the 1,783 authors here whose files we hold, read 2026-08-07

  • Ask questions one at a timein 81 of 1048, across 64 files
  • Provide a recommended answer for each questionin 73 of 1048, across 50 files
  • Explore the codebase instead of asking answerable questionsin 66 of 1048, across 42 files
  • Resolve dependencies between decisions one-by-onein 42 of 1048, across 17 files
  • Interview the user relentlessly about the planin 38 of 1048, across 13 files
  • Order findings by severityhere, and in 31 of 1048
  • Resolve each branch of the decision treein 27 of 1048, across 5 files
  • Run a grilling sessionin 26 of 1048, across 5 files
  • Update CONTEXT.md immediately when a term is resolvedin 26 of 1048, across 11 files
  • Propose precise canonical terms for vague languagein 25 of 1048, across 7 files
  • Create documentation files lazilyin 24 of 1048, across 5 files
  • Assign severity to every findingin 24 of 1048

Said here and by no other author read

  • read bootstrap files in exact order
  • resolve effective spec chain before reviewing
  • identify nearest spec granting write authority before editing
  • compare diff against applicable spec contract entries
  • check whether verification matches behavioral risk
  • include file, contract, risk, and likely fix per finding

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,144. 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.