Agent skill

final-verification

SAM Stage 7 — Goal-backward certification that the feature achieves its original objectives. Used when all tasks pass forensic review; starts from expected outcomes, works backwards to verify each was achieved. Returns CERTIFIED or NOT_CERTIFIED with specific gaps.

Stars 33
Forks 4

Install this agent skill to your Project

npx add-skill https://github.com/Jamie-BitFlight/claude_skills/tree/main/plugins/development-harness/skills/final-verification

SKILL.md

SAM Stage 7 — Final Verification

Role

You are the final verification certifier for the SAM pipeline. You determine whether the implemented feature achieves the original objectives defined in Stage 1 Discovery. You work backwards from goals to evidence.

Core Principle

Goal-backward verification. Do not start from what was built and ask "is this good enough?" Start from what SHOULD be true and verify it IS true. This prevents anchoring bias from the implementation details.

When to Use

  • After all tasks have passed Stage 6 Forensic Review with COMPLETE verdicts
  • As the final gate before declaring the feature ready for commit or PR
  • When re-certifying after NOT_CERTIFIED gaps are addressed

Process

mermaid
flowchart TD
    Start([All reviews COMPLETE + feature-context + architect artifacts]) --> G1[1. Extract original goals]
    G1 --> G2[2. For each goal — identify required truths]
    G2 --> G3[3. For each truth — verify in codebase]
    G3 --> G4[4. Check acceptance criteria from PLAN]
    G4 --> G5[5. Run quality gates]
    G5 --> Decide{All goals verified with evidence?}
    Decide -->|Yes| Certified[CERTIFIED]
    Decide -->|No| NotCertified[NOT_CERTIFIED]
    Certified --> Done([ARTIFACT:VERIFICATION])
    NotCertified --> Gap[Identify gaps — create new tasks]
    Gap --> Loop[Loop to Stage 4 for new tasks]
    Loop --> Done

Step 1 — Extract Original Goals

Read the feature-context artifact via artifact_read(issue_number={issue}, artifact_type="feature-context") and extract:

  • All goals from the Goals section
  • All anti-goals from the Anti-Goals section
  • All functional requirements
  • All non-functional requirements

These are the ONLY criteria for certification. Features not in the original discovery are out of scope for this verification.

Step 2 — Identify Required Truths

For each goal, enumerate what must be TRUE in the codebase:

text
Goal: "Users can authenticate via OAuth2"
Required truths:
  - OAuth2 client configuration exists
  - Authentication endpoint handles OAuth2 flow
  - Token validation middleware is integrated
  - Error cases return appropriate HTTP status codes
  - Session management stores authenticated state

Step 3 — Verify Each Truth

For each required truth, verify it through direct observation:

  • Read the relevant files and confirm the implementation exists
  • Run the relevant tests and confirm they pass
  • Check integration points and confirm they connect
  • Verify anti-goals are NOT violated (no scope creep)

Document evidence for each truth — file paths, test output, observed behavior.

Step 4 — Check Acceptance Tests

Read the architect artifact via artifact_read(issue_number={issue}, artifact_type="architect") and extract the acceptance tests (Given/When/Then). For each acceptance test:

  • Verify the precondition (Given) can be established
  • Verify the action (When) is possible
  • Verify the outcome (Then) is observable and correct

Step 5 — Run Quality Gates

Run the project's quality gates to confirm the entire feature passes:

  • Format check
  • Lint check
  • Type check (if applicable)
  • Full test suite
  • Any project-specific gates from the language manifest

For the quality gate protocol, reference /dh:validation-protocol.

Input

  • All review results via sam_read(plan="{plan_id}", task="{task_id}") per task — review content is stored in task body sections
  • Feature-context artifact via artifact_read(issue_number={issue}, artifact_type="feature-context")
  • Architect artifact via artifact_read(issue_number={issue}, artifact_type="architect")
  • Read access to the codebase

Output

Append to the plan via sam_update(address="{plan_id}/{task_id}", append_section="Final Verification", section_content="{verification_markdown}") where {verification_markdown} follows this template:

markdown
# ARTIFACT:VERIFICATION

## Verdict

<CERTIFIED / NOT_CERTIFIED>

## Feature

<feature name from DISCOVERY>

## Goal Verification

### Goal 1 — <goal text>

| Required Truth | Verified | Evidence |
|---------------|----------|----------|
| <what must be true> | YES / NO | <file path, test output, observation> |

### Goal 2 — <goal text>

| Required Truth | Verified | Evidence |
|---------------|----------|----------|
| <what must be true> | YES / NO | <file path, test output, observation> |

## Anti-Goal Compliance

| Anti-Goal | Violated | Evidence |
|-----------|----------|----------|
| <what must NOT happen> | NO / YES | <observation confirming compliance or violation> |

## Acceptance Test Results

| Test | Given | When | Then | Result |
|------|-------|------|------|--------|
| <test name> | <precondition> | <action> | <expected outcome> | PASS / FAIL |

## Quality Gates

| Gate | Result | Output |
|------|--------|--------|
| Format | PASS / FAIL | <summary> |
| Lint | PASS / FAIL | <summary> |
| Typecheck | PASS / FAIL | <summary> |
| Tests | PASS / FAIL | <summary> |

## NFR Verification

| NFR | Criterion | Verified | Evidence |
|-----|-----------|----------|----------|
| <from DISCOVERY> | <measurable target> | YES / NO | <measurement or observation> |

## Gaps (if NOT_CERTIFIED)

1. **<gap title>** — <what goal is unmet, what truth is false, what evidence is missing>

## Remediation Path (if NOT_CERTIFIED)

New tasks to create — loop back to Stage 4 (Task Decomposition):

1. **<task title>** — <what must be done to close the gap>

## Certification Statement (if CERTIFIED)

All goals from ARTIFACT:DISCOVERY are verified with evidence.
All acceptance tests from ARTIFACT:PLAN pass.
All quality gates pass.
No anti-goals are violated.
Feature is ready for commit/PR.

NOT_CERTIFIED Loop

mermaid
flowchart TD
    NotCert([NOT_CERTIFIED]) --> Gaps[Document specific gaps]
    Gaps --> NewTasks[Create new TASK files for gaps]
    NewTasks --> Stage4[Stage 4 — Decompose gap tasks]
    Stage4 --> Stage5[Stage 5 — Execute gap tasks]
    Stage5 --> Stage6[Stage 6 — Review gap executions]
    Stage6 --> Stage7[Stage 7 — Re-certify]
    Stage7 --> Q{CERTIFIED?}
    Q -->|Yes| Done([Feature complete])
    Q -->|No| Gaps

Behavioral Rules

  • Always start from goals and work backward — never start from implementation
  • Verify anti-goals explicitly — absence of violation must be confirmed
  • Do not add requirements not in ARTIFACT:DISCOVERY
  • Every verification must cite evidence (file path, command output, observation)
  • NFRs must be measured, not assumed ("latency < 200ms" requires a measurement)
  • CERTIFIED requires ALL goals verified — partial certification does not exist

Success Criteria

  • Every goal from DISCOVERY verified with evidence
  • Every anti-goal confirmed not violated
  • Every acceptance test from PLAN passes
  • All quality gates pass
  • All NFRs measured and within thresholds
  • Certification statement (or gap list) is complete and evidence-based

Expand your agent's capabilities with these related and highly-rated skills.

Jamie-BitFlight/claude_skills

ccc

This skill should be used when code search is needed (whether explicitly requested or as part of completing a task), when indexing the codebase after changes, or when the user asks about ccc, cocoindex-code, or the codebase index. Trigger phrases include 'search the codebase', 'find code related to', 'update the index', 'ccc', 'cocoindex-code'.

33 4
Explore
Jamie-BitFlight/claude_skills

agent-browser

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

33 4
Explore
Jamie-BitFlight/claude_skills

delegate

Quick delegation template for sub-agent prompts. Use when assigning work to a sub-agent, before invoking the Agent tool, or when preparing prompts for specialized agents. Provides the WHERE-WHAT-WHY framework. For comprehensive delegation guidance, activate the agent-orchestration how-to-delegate skill.

33 4
Explore
Jamie-BitFlight/claude_skills

swarm-spawning

Spawn agents and teammates in Claude Code swarms. Use when choosing between subagents vs teammates, selecting agent types (Explore, Plan, general-purpose, plugin agents), configuring spawn backends (in-process, tmux, iterm2), or setting environment variables for spawned agents.

33 4
Explore
Jamie-BitFlight/claude_skills

knowledge-explorer

Manage the research/ knowledge base (KB) of tool and library research entries. Use when browsing KB topics, adding new research entries, updating existing entries with dated revisions, fetching GitHub repo metadata into a draft KB entry, or migrating old-format entries to skill-spec frontmatter. Triggers on tasks like "what do we have on X", "add this to the KB", "update the KB entry for Y", "fetch github info for owner/repo", or "migrate old entries".

33 4
Explore
Jamie-BitFlight/claude_skills

design-anti-patterns

Enforce anti-AI UI design rules based on the Uncodixfy methodology. Use when generating HTML, CSS, React, Vue, Svelte, or any frontend UI code. Prevents "Codex UI" — the generic AI aesthetic of soft gradients, floating panels, oversized rounded corners, glassmorphism, hero sections in dashboards, and decorative copy. Applies constraints from Linear/Raycast/Stripe/GitHub design philosophy: functional, honest, human-designed interfaces. Triggers on: UI generation, dashboard building, frontend component creation, CSS styling, landing page design, or any task producing visual interface code.

33 4
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results