completion-validation
Reconciles intended feature against blind inference.
Skill by autoencoders
Chalk Labs
Need help deploying completion-validation?
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/autoencoders-completion-validation
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: testing, data, ml
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/completion-validation
What's inside
- Skill name
completion-validation- Description
- Reconciles intended feature against blind inference. Categorical alignment (MATCH/PARTIAL/MISMATCH) with critical miss override. No testing — comparison and decision only. Triggers on "validate completion", "reconcile intent vs implementation", "is feature complete", or as final step in feature pipeline.
- License
- CC-BY-4.0
- Instructions
- 268 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Completion Validation Compare what was intended vs. what blind verification inferred. Make a decision. No sugarcoating. ## Execution Modes | Mode | Behavior | |------|----------| | `full` | Full reconciliation | | `fast` | Full reconciliation | | `prototype` | Full reconciliation | | `spike` | Skip this skill entirely | In `spike` mode, validation is skipped. The user owns the consequences. ## What This Is (and Isn't) **IS:** Reconciliation — compare two descriptions, decide accept/iterate/reject **IS NOT:** - A test suite - A second QA pass - A re-verification of code If testing is needed, the decision is **Iterate** and testing happens back in implementation. ## Inputs Required 1. **Original request** (from evaluator or requester) — 1 sentence summary 2. **Blind inference** (from implementation-verification) — the inference snapshot 3. **Critical criteria** (from planning) — what MUST be present ## Output Format (Fixed, Minimal) ``` RECONCILIATION DECISION ═══════════════════════ Original request (1 sentence): [What was supposed to be built] Blind inference (1 sentence): [What implementation-verification says was built] --- Alignment: [MATCH / PARTIAL / MISMATCH] Critical miss: [Yes / No] If Yes: [What critical item is missing — 1 bullet] Confidence: [High / Medium / Low] (from blind inference) If Low + observable intent: [Flag — implementation should be more legible] Gaps (0-3 bullets): - [Gap 1] - [Gap 2] --- Decision: [Accept / Accept with caveat / Iterate / Reject] If Iterate: Smallest fix: [1 bullet — the ONE thing to change] Updated criterion: [1 bullet — how to verify it's fixed] Iteration: [N] of 2 max If Accept with caveat: Caveat: [What's not fully verified and why that's okay] If Reject: Reason: [Why this can't be fixed with small iterations] ``` ## Decision Logic (Explicit Precedence) Evaluate in this order. Stop at first applicable rule. ``` ┌─────────────────────────────────────────────────────────────┐ │ PRECEDENCE ORDER │ ├─────────────────────────────────────────────────────────────┤ │ │ │ 1. CRITICAL MISS CHECK (first — overrides everything) │ │ ┌─────────────────────────────────────────────────┐ │ │ │ Critical miss = Yes? │ │ │ │ → Force Alignment = PARTIAL │ │ │ │ → Decision = Iterate (critical fix) │ │ │ │ → Stop here │ │ │ └─────────────────────────────────────────────────┘ │ │ │ │ │ ▼ No │ │ 2. CONFIDENCE CHECK (second — routes weak evidence) │ │ ┌─────────────────────────────────────────────────┐ │ │ │ Confidence = Low? │ │ │ │ If intent is observable: │ │ │ │ → Decision = Iterate (improve legibility) │ │ │ │ → Stop here │ │ │ │ If intent is non-observable: │ │ │ │ → Decision = Accept with caveat │ │ │ │ OR escalate verification method │ │ │ │ → Stop here │ │ │ └─────────────────────────────────────────────────┘ │ │ │ │ │ ▼ Medium/High │ │ 3. ALIGNMENT CHECK (third — normal path) │ │ ┌─────────────────────────────────────────────────┐ │ │ │ MATCH → Accept │ │ │ │ PARTIAL → Iterate (smallest fix) │ │ │ │ MISMATCH → Reject │ │ │ └─────────────────────────────────────────────────┘ │ │ │ └─────────────────────────────────────────────────────────────┘ ``` ## Decision Definitions ### Accept All critical items present. Blind inference captures intent. Ship it.
From autoencoders/skills, CC-BY-4.0.
More from autoencoders/skills
- 0.414score
feature-planning
autoencoders
Creates detailed implementation plans for approved features.
SkillDFreshness: DClaude Code +9 - 0.414score
feature-evaluator
autoencoders
Structured evaluation of proposed features for existing projects.
SkillDFreshness: DClaude Code +9 - 0.414score
feature-implementation
autoencoders
Executes implementation plans with closed-loop verification.
SkillDFreshness: DClaude Code +9 - 0.414score
scaffold
autoencoders
Unified skill for Python project scaffolding with uv workspaces.
SkillDFreshness: DClaude Code +9 - 0.409score
implement
autoencoders
Entry point orchestrator for the feature development pipeline.
SkillDFreshness: DClaude Code +9 - 0.409score
implementation-verification
autoencoders
Blind inference of what was implemented from code changes only.
SkillDFreshness: DClaude Code +9 - 0.396score
hello
autoencoders
A simple test skill that says hi.
SkillDFreshness: DClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Feb 6, 2026 (9 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- CC-BY-4.0
- Repository
- github.com/autoencoders/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from autoencoders
- 0.414score
feature-planning
autoencoders
Creates detailed implementation plans for approved features.
SkillDFreshness: DClaude Code +9 - 0.414score
feature-evaluator
autoencoders
Structured evaluation of proposed features for existing projects.
SkillDFreshness: DClaude Code +9 - 0.414score
feature-implementation
autoencoders
Executes implementation plans with closed-loop verification.
SkillDFreshness: DClaude Code +9 - 0.414score
scaffold
autoencoders
Unified skill for Python project scaffolding with uv workspaces.
SkillDFreshness: DClaude Code +9 - 0.409score
implement
autoencoders
Entry point orchestrator for the feature development pipeline.
SkillDFreshness: DClaude Code +9 - 0.409score
implementation-verification
autoencoders
Blind inference of what was implemented from code changes only.
SkillDFreshness: DClaude 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 completion-validation 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
completion-validation skill. Agent Mart. https://agentmart.dev/skills/autoencoders-completion-validation (last updated 2026-10-24)