Skip to content
Back to skills

Subagents Validation

ASecurity

Validate persona, skill, rules, stack-pack, catalog, and runtime-surface consistency.

  • 11 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 23, 2026
ai-agents

Security analysis

A100/100

Scanned October 6, 2026

npx -y skills add startmeupai/swe-agents --skill subagents-validation --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Subagents Validation?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Subagents Validation
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/startmeupai-subagents-validation/badge)](https://www.skillsdirectory.com/skills/startmeupai-subagents-validation)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: subagents-validation
description: Validate persona, skill, rules, stack-pack, catalog, and runtime-surface consistency.
---

# Subagents Validation

## Trigger Conditions

Use after agent-system changes or before publishing agent infrastructure.

## Required Inputs

- Core and stack-pack persona, skill, rules-fragment, and manifest locations,
  plus the catalog, routing guide, check scripts, and runtime-registration
  locations.

## Workflow

1. Validate directories, filenames, pack manifests, and YAML frontmatter.
2. Confirm personas describe ownership rather than long procedures.
3. Confirm skills contain triggers, bounded workflows, checks, evidence, and completion.
4. Resolve every persona-to-skill and catalog reference; a pack persona may use
   only core skills, its own pack's skills, or skills from a pack it requires.
5. Confirm core content names no pack persona or skill, and that agent and
   skill names are unique across core and every pack.
6. Check contradictions, sanitization, stale generated copies, and missing
   runtime surfaces.
7. Simulate representative prompts and record dispatch ambiguity.

## Deterministic Checks

- In the reference repository, run `pnpm check:agents`, `pnpm check:skills`,
  `pnpm check:references`, and `pnpm check:stacks`. In an installed project,
  run the agent-system checks it provides and record any that are unavailable.

## Safety and Permission Boundaries

- Validation is read-only unless the user separately authorizes corrections.

## Required Evidence

- Passed checks, precise violations, simulated prompts, and untested runtime compatibility.

## Completion Condition

- Deterministic checks pass and all remaining compatibility questions are explicit.

## Example

`Validate this reference's core and pack agents, skills, rules, and links.`

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…