Adversarial security and resilience analysis — auto-triggered during /review and /test based on task classification. Provides attack surface analysis, boundary testing, auth bypass attempts, dependency chain attacks, and Beast Mode stress testing.

2026-04-1692

dispatching-parallel-agents

KbWen/agentic-os

Evaluate when to dispatch parallel agents; use a 4-step pattern to split, coordinate, and integrate results.

2026-04-1292

name	verification-before-completion
description	Enforce "no evidence = no completion"; run Gate Function verification before declaring done.

Verification Before Completion

Overview

Any "completed" statement MUST be supported by evidence. This skill provides a pre-completion Gate Function, ensuring results are reproducible, traceable, and deliverable.

Ironclad Rules

No evidence = no completion claim.
Evidence MUST be reproducible by others; verbal descriptions are NOT accepted.
If tests/checks fail, the status MUST revert to in-progress.

When to Use

Tasks are preparing to enter /ship.
Before submitting PRs or handing off work.
After completing high-risk changes (data, permissions, core flows).

Gate Function

Scope Gate: Confirm changes only cover the agreed scope.
Quality Gate: Execute required tests and static checks.
Evidence Gate: Compile reproducible evidence (commands, outputs, versions).
Risk Gate: Confirm rollback strategies and known risks.
Communication Gate: Output a completion summary (changes, validation, constraints).

Minimum Evidence Checklist

At least one set of tests/checks directly related to this change.
Specific execution commands and results (Success/Failure/Warning).
If there are unfinished items, explicitly mark them with follow-up suggestions.

Common Mistakes

Claiming "it should work" without any output evidence.
Testing only the happy path, missing regressions or edge cases.
Treating old test results as evidence for new changes.