verify
Human-facing skill for reviewing a built feature against its feature doc, decomposition, and definition of done, then saving a verification report with…
Skill by tykowale
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/tykowale-verify
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: documents
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. Checked Oct 25, 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/verify
What's inside
- Skill name
verify- Description
- Human-facing skill for reviewing a built feature against its feature doc, decomposition, and definition of done, then saving a verification report with recommended next improvements.
- License
- MIT
- Instructions
- 66 lines in SKILL.md · view on GitHub
Files (3)
- SKILL.md
- feature-review-checklist.md · resources/
- verification-report-template.md · resources/
SKILL.md (first lines)
# Verify This is a human-facing skill. Use it when the user wants Darwin to inspect what has already been built and decide whether the feature is actually complete, what is missing, and what should improve next. ## Purpose Turn the current built state of a feature into a saved verification report. The report should answer: - what source docs define the feature - what slices appear complete versus incomplete - what was verified with fresh evidence - what is missing, weak, or inconsistent - whether the feature feels good enough as-is - what the next improvements should be, if any ## Dependencies - `using-darwin` - `clarifying-intent` - `verification-before-completion` - `debrief-and-learn` - `skills/verify/resources/feature-review-checklist.md` - `skills/verify/resources/verification-report-template.md` ## Clarification Rule Ask at most 3 clarifying questions before starting the review. Questions should focus on: - which feature doc or decomposition doc is the source of truth - what "done" means for this review - whether the user wants correctness-only review or also UX/product improvement suggestions ## Workflow 1. Read the relevant source artifacts: feature doc, decomposition, and any definition-of-done notes. 2. Inspect the current implementation and identify what appears to be shipped. 3. Use the feature review checklist to compare the built state against the intended scope and success criteria. 4. Run fresh verification using `verification-before-completion` for the claims the report makes. 5. Save the result to `docs/verification/<date>-<slug>.md` using the verification report template. 6. Run a collaborative debrief with the operator about the quality of the feature, the review process, and whether Darwin missed anything. 7. Canonize any clear Darwin improvement immediately when it belongs in a skill, prompt, or resource. 8. End with the next best step: feature is solid as-is, build the next slice, fix a gap, or redefine the scope. ## Output To The User Keep the interaction simple: - a few high-value questions only when needed - a concise readout of what is solid and what is missing - a saved verification report in `docs/verification/` - a clear recommendation for the next step ## Guardrails - Do not confuse code inspection with verification. Claims still need fresh evidence. - Do not invent problems when the feature is genuinely in a good place. - Do not claim completeness just because the code exists. - Do not skip UX or product-level gaps when the user asked for a fuller review.
From tykowale/darwin, MIT.
More from tykowale/darwin
- 0.406score
build
tykowale
Human-facing skill for turning a build request into an approved plan, a TDD execution loop, and a debrief that improves Darwin.
SkillDFreshness: DClaude Code +9 - 0.406score
execution-coordination
tykowale
Internal meta skill for dispatching subagents, coordinating execution, and integrating reviewed results back into the main flow.
SkillDFreshness: DClaude Code +9 - 0.406score
frontend-work
tykowale
Internal meta skill for routing frontend implementation through reusable UI, UX, and interaction resources without widening Darwin's human API.
SkillDFreshness: DClaude Code +9 - 0.406score
debrief-and-learn
tykowale
Internal meta skill for reflecting on completed work, extracting lessons, and feeding them back into Darwin's self-improvement loop.
SkillDFreshness: DClaude Code +9 - 0.406score
using-darwin
tykowale
Internal meta skill for starting work in Darwin.
SkillDFreshness: DClaude Code +9 - 0.406score
define
tykowale
Human-facing skill for clarifying a potential feature, pressure-testing it, and producing a saved feature document before implementation planning.
SkillDFreshness: DClaude Code +9 - 0.404score
verification-before-completion
tykowale
Internal meta skill for verifying work with fresh evidence before claiming success, committing, or pushing.
SkillDFreshness: DClaude Code +9 - 0.404score
writing-plans
tykowale
Internal meta skill for turning an approved design into an execution-ready implementation plan.
SkillDFreshness: DClaude Code +9 - 0.404score
clarifying-intent
tykowale
Internal meta skill for resolving purpose, constraints, and success criteria before design and implementation planning.
SkillDFreshness: DClaude Code +9 - 0.404score
create
tykowale
Human-facing skill for creating a new Darwin skill or adding a nested resource inside an existing skill.
SkillDFreshness: DClaude Code +9 - 0.404score
decompose
tykowale
Human-facing skill for turning a feature document into modular vertical slices and incremental delivery steps.
SkillDFreshness: DClaude Code +9 - 0.404score
delete
tykowale
Human-facing skill for deleting a Darwin skill or removing nested files from within a skill directory.
SkillDFreshness: DClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Apr 5, 2026 (6 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/tykowale/darwin
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from tykowale
- 0.406score
define
tykowale
Human-facing skill for clarifying a potential feature, pressure-testing it, and producing a saved feature document before implementation planning.
SkillDFreshness: DClaude Code +9 - 0.406score
build
tykowale
Human-facing skill for turning a build request into an approved plan, a TDD execution loop, and a debrief that improves Darwin.
SkillDFreshness: DClaude Code +9 - 0.406score
debrief-and-learn
tykowale
Internal meta skill for reflecting on completed work, extracting lessons, and feeding them back into Darwin's self-improvement loop.
SkillDFreshness: DClaude Code +9 - 0.406score
execution-coordination
tykowale
Internal meta skill for dispatching subagents, coordinating execution, and integrating reviewed results back into the main flow.
SkillDFreshness: DClaude Code +9 - 0.406score
frontend-work
tykowale
Internal meta skill for routing frontend implementation through reusable UI, UX, and interaction resources without widening Darwin's human API.
SkillDFreshness: DClaude Code +9 - 0.406score
using-darwin
tykowale
Internal meta skill for starting work in Darwin.
SkillDFreshness: DClaude Code +9
Cite this page
verify skill. Agent Mart. https://agentmart.dev/skills/tykowale-verify (last updated 2026-10-24)