| name | vibeguard |
| description | Lightweight anti-hallucination workflow for task kickoff, review prioritization, and regression retrospectives. Use when the user asks for guardrails, task contracts, risk scoring, or review templates. |
VibeGuard
Lightweight VibeGuard for everyday use inside Claude Arsenal.
This skill helps with three moments where AI-assisted work usually drifts:
- before implementation starts
- while review findings are being prioritized
- after regressions or avoidable mistakes happen
What This Skill Includes
- A task contract checklist for scoping work before coding
- A scoring matrix for prioritizing findings with evidence
- A weekly review template for regression and guardrail retrospectives
Hard Boundaries
This skill is intentionally lightweight.
It does:
- guide task kickoff
- structure review output
- make risk tradeoffs explicit
It does not:
- install hooks
- run guard scripts automatically
- patch global Claude or Codex configuration
- claim that full VibeGuard enforcement is active
If the user asks for automated interception, repo-level rules, or environment setup, escalate to the full VibeGuard repository and tooling.
When To Use It
Trigger this skill when the user asks for:
- anti-hallucination guardrails
- task startup checks
- review scoring or triage
- regression prevention
- retrospective templates
- "vibeguard" by name
Workflow
1. Start With The Task Contract
Open references/task-contract.yaml and confirm:
- goal
- source of truth
- acceptance criteria
- scope
Do not move into implementation until these are concrete.
2. Score Findings Before Acting
Open references/scoring-matrix.md and score each finding on:
- impact
- effort
- risk
- confidence
Use the score to separate urgent fixes from weak guesses.
3. Pick The Right Delivery Mode
- 1-2 files: implement directly once the contract is clear
- 3-5 files: write a short spec or step plan first
- 6+ files: produce a full design/spec and staged execution plan
4. Review The Failure, Not Just The Symptom
If something regressed, identify which defense failed:
- scope control
- source-of-truth validation
- verification depth
- review prioritization
- missing guardrail
Then capture the follow-up in references/review-template.md.
Expected Outputs
Depending on the request, produce one of these:
- a filled task contract
- a scored findings table with priorities
- a short regression review or weekly retrospective
Reference Files
