Skip to content
Back to skills

Crew

ASecurity

Verify the crew:explain skill translates jargon-heavy gate findings and phase summaries into grade-8 plain language

  • 9 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 18, 2026
ai-agentspythongobashtesting

Security analysis

A100/100

Pro scans all 21 files and shows the line behind each finding

Scanned October 18, 2026

npx -y skills add mikeparcewski/wicked-garden --skill crew --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Crew?

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

Security grade badge for Crew
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/mikeparcewski-crew/badge)](https://www.skillsdirectory.com/skills/mikeparcewski-crew)

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: plain-language-translation-skill
title: Plain-Language Translation Skill (skills/crew/explain/SKILL.md)
description: Verify the crew:explain skill translates jargon-heavy gate findings and phase summaries into grade-8 plain language
type: testing
difficulty: beginner
estimated_minutes: 8
covers:
  - "#519 — plain-language translation skill acceptance criteria"
  - skills/crew/explain/SKILL.md
  - Plain: line convention in clarify/objective.md, design.md
ac_ref: "skills/crew/explain/SKILL.md — crew:explain skill"
---

# Plain-Language Translation Skill

@manual — The `Run:` steps that invoke `/wicked-garden:crew:explain` directly require
an active Claude session. The structural validation steps (file existence, convention
checks) are harness-runnable.

This scenario validates `skills/crew/explain/SKILL.md`:

1. Skill file exists and is ≤ 200 lines (Tier-2 size budget).
2. Skill defines output rules: 2-4 sentences, grade-8 reading level, no jargon.
3. Input jargon block → output contains a `**Plain:**` line in `paired` mode.
4. Cross-check: `**Plain:**` convention is present in clarify/objective.md and design.md
   phase templates (if they exist).

## Setup

```bash
export PLUGIN_ROOT="${CLAUDE_PLUGIN_ROOT}"
```

---

## Case 1: Skill file exists and is within size budget

```bash
Run: test -f "${PLUGIN_ROOT}/skills/crew/explain/SKILL.md" && echo "PASS: explain/SKILL.md exists"
Assert: PASS: explain/SKILL.md exists
```

```bash
Run: sh "${PLUGIN_ROOT}/scripts/_python.sh" -c "
import pathlib
skill = pathlib.Path('${PLUGIN_ROOT}/skills/crew/explain/SKILL.md')
lines = skill.read_text().splitlines()
assert len(lines) <= 200, f'SKILL.md exceeds 200-line budget: {len(lines)} lines'
print(f'PASS: explain/SKILL.md is {len(lines)} lines (within 200-line budget)')
"
Assert: PASS: explain/SKILL.md is N lines (within 200-line budget)
```

---

## Case 2: Skill defines output rules — grade-8, 2-4 sentences, no jargon

```bash
Run: sh "${PLUGIN_ROOT}/scripts/_python.sh" -c "
import pathlib
content = pathlib.Path('${PLUGIN_ROOT}/skills/crew/explain/SKILL.md').read_text()

required_phrases = [
    'grade',       # grade-8 reading level referenced
    'jargon',      # no specialist vocab rule
    'Plain',       # Plain: line convention
]
for phrase in required_phrases:
    assert phrase.lower() in content.lower(), f'Missing required phrase: {phrase!r}'

print('PASS: explain/SKILL.md contains grade-level, jargon, and Plain: convention references')
"
Assert: PASS: explain/SKILL.md contains grade-level, jargon, and Plain: convention references
```

---

## Case 3: REJECT jargon → plain language — no specialist vocab in output

This test simulates the skill's translation rule by checking the prohibition list.
The actual invocation is @manual (requires live session).

```bash
Run: sh "${PLUGIN_ROOT}/scripts/_python.sh" -c "
# Simulate: input jargon block → expected plain output characteristics
JARGON_INPUT = '''
Gate: design-quality
Verdict: CONDITIONAL
Score: 0.65 / 0.70 (BLEND: 0.4*min + 0.6*avg)
Conditions:
  - FR-3 blast radius not quantified in architecture.md
  - parallelization_check missing from executor-status.json
  Dispatch mode: parallel (3 reviewers in council)
'''

# What explain skill MUST NOT include in output:
BANNED_OUTPUT_TERMS = [
    'CONDITIONAL', 'BLEND', 'blast radius', 'parallelization_check',
    'executor-status', 'dispatch mode', 'council', 'per_reviewer'
]

# What it SHOULD include (plain equivalents)
REQUIRED_CONCEPTS = [
    # At least one of these plain-language concepts must appear
    'fix', 'blocked', 'needs', 'missing', 'reviewers', 'next'
]

# Verify the skill SKILL.md documents these substitution rules
content = open('${PLUGIN_ROOT}/skills/crew/explain/SKILL.md').read()
for banned in BANNED_OUTPUT_TERMS[:3]:  # check first 3 are mentioned as banned
    assert banned in content or banned.lower() in content.lower(), (
        f'Expected banned term {banned!r} to be listed in skill output rules'
    )

print('PASS: skill documents jargon terms that must be substituted in output')
print('  (full translation test is @manual — requires live session)')
"
Assert: PASS: skill documents jargon terms that must be substituted in output
```

---

## Case 4: **Plain:** line convention in clarify/objective.md template

The `explain` skill produces a `**Plain:**` line in `paired` mode. Verify that
crew phase output templates (if present) use this convention.

```bash
Run: sh "${PLUGIN_ROOT}/scripts/_python.sh" -c "
import pathlib, sys
plugin_root = pathlib.Path('${PLUGIN_ROOT}')
# Check skills/crew/explain/SKILL.md for the paired-mode Plain: convention
skill_content = (plugin_root / 'skills/crew/explain/SKILL.md').read_text()

assert '**Plain:**' in skill_content or 'Plain:' in skill_content, (
    'SKILL.md does not document the **Plain:** line convention'
)
print('PASS: **Plain:** convention documented in explain/SKILL.md')

# Check if any phase agent files reference the Plain: convention
agents_dir = plugin_root / 'agents/crew'
plain_refs = []
for f in agents_dir.glob('*.md'):
    if 'Plain' in f.read_text():
        plain_refs.append(f.name)

if plain_refs:
    print(f'PASS: Plain: convention also referenced in {len(plain_refs)} crew agent(s): {plain_refs}')
else:
    print('INFO: Plain: convention not found in crew agents (may be skills-only)')
"
Assert: PASS: **Plain:** convention documented in explain/SKILL.md
```

---

## Case 5: crew:explain command exists and is discoverable

```bash
Run: test -f "${PLUGIN_ROOT}/commands/crew/explain.md" && echo "PASS: commands/crew/explain.md exists" || echo "INFO: explain.md not in commands (may be skills-only — check skill registration)"
Assert: PASS or INFO (skill must be reachable via /wicked-garden:crew:explain)
```

```bash
Run: sh "${PLUGIN_ROOT}/scripts/_python.sh" -c "
import pathlib
# Verify explain skill is registered in skills/crew/explain/
skill_dir = pathlib.Path('${PLUGIN_ROOT}/skills/crew/explain')
assert skill_dir.is_dir(), f'explain skill directory missing at {skill_dir}'
assert (skill_dir / 'SKILL.md').exists(), 'SKILL.md missing in explain/'
print('PASS: skills/crew/explain/SKILL.md registered and discoverable')
"
Assert: PASS: skills/crew/explain/SKILL.md registered and discoverable
```

---

## Success Criteria

- [ ] `skills/crew/explain/SKILL.md` exists and is ≤ 200 lines
- [ ] SKILL.md documents: grade-8 reading level, no jargon, Plain: convention
- [ ] SKILL.md lists banned output terms with plain-language substitutions
- [ ] `**Plain:**` line convention is documented in SKILL.md
- [ ] `commands/crew/explain.md` exists OR skill is reachable as `/wicked-garden:crew:explain`
- [ ] (@manual) Input jargon block → output 2-4 sentences with no banned specialist terms

Files in this skill

  • 01-end-to-end-feature.md9.6 KB
  • 01-memory-storage.md5.5 KB
  • 02-autonomous-completion.md7.4 KB
  • 02-onboarding.md7.1 KB
  • 03-crew-detection.md7.7 KB
  • 04-jam-triggers.md8.4 KB
  • 05-multi-project-management.md10.4 KB
  • 05-qe-gate.md12.9 KB
  • 07-autonomous-assumptions.md5.5 KB
  • affected-repos-advisory.md9.2 KB
  • artifact-state-machine.md8.9 KB
  • blind-reviewer-multi-reviewer-invariant.md7.7 KB
  • bus-cutover-dispatch-log-chain-id-uniqueness.md6.5 KB
  • bus-cutover-gate-result-security-floor.md48.6 KB
  • bus-emit-lint-blocks-orphan-write.md8.1 KB
  • challenge-gate-enforcement.md7.3 KB
  • consensus-chain-id-uniqueness.md7.7 KB
  • context-isolation.md9.3 KB
  • contrarian-convergence-collapse.md4.7 KB
  • contrarian-missing-section.md3.2 KB

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…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.