Skip to content
Back to skills

2917 23 Pre Commit Config Update 31fdb907

ASecurity

**Status**: NOT STARTED **Agent**: @python-pytest-architect **Dependencies**: Task 11 **Priority**: 4 **Complexity**: M **Accuracy Risk**: Low

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
testingpython

Works with

  • cli
  • mcp

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 2917-23-pre-commit-config-update_31fdb907 --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 2917 23 Pre Commit Config Update 31fdb907?

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

Security grade badge for 2917 23 Pre Commit Config Update 31fdb907
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-2917-23-pre-commit-config-update-31fdb907/badge)](https://www.skillsdirectory.com/skills/tools-only-2917-23-pre-commit-config-update-31fdb907)

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

SKILL.md
---
task: "23"
title: "Pre-Commit Config Update"
status: not-started
agent: "@python-cli-architect"
dependencies: ["3"]
priority: 5
complexity: s
---

## Task 18: AgentEnumValidator Tests

**Status**: NOT STARTED
**Agent**: @python-pytest-architect
**Dependencies**: Task 11
**Priority**: 4
**Complexity**: M
**Accuracy Risk**: Low

#### Context

AgentEnumValidator needs comprehensive test coverage for enum validation and suggestion quality.

#### Objective

Create complete test suite for AgentEnumValidator achieving 90%+ coverage.

#### Required Inputs

- Architecture spec: ./architect-plugin-linter.md lines 983-995 (AgentEnumValidator test requirements)
- Completed Task 11: AgentEnumValidator implementation
- Test pattern: plugins/plugin-creator/tests/test_frontmatter_validator.py

#### Requirements

1. Create test_agent_enum_validator.py with pytest test class
2. Test valid agent enum combinations (parametrized)
3. Test invalid model values → AG001 with suggestions
4. Test invalid permissionMode values → AG002 with suggestions
5. Test invalid memory values → AG003 with suggestions
6. Test missing required fields → AG004, AG005
7. Test invalid maxTurns → AG006
8. Test invalid color values → AG007 (warning)
9. Test disallowedTools validation → AG008 (warning)
10. Test tools/disallowedTools conflict → AG009
11. Test hooks reference validation → AG010
12. Parametrize valid enum combination tests
13. Verify suggestion quality for invalid enum values
14. Achieve 90%+ coverage

#### Constraints

- MUST verify enum value suggestions are accurate
- MUST test all valid enum combinations
- MUST test file type filtering (only runs on AGENT files)
- MUST achieve minimum 90% coverage

#### Expected Outputs

- New file: `plugins/plugin-creator/tests/test_agent_enum_validator.py` (~250 lines)
- Test fixtures for valid/invalid agent frontmatter
- Parametrized tests for enum combinations
- Coverage report showing 90%+ for AgentEnumValidator

#### Acceptance Criteria

1. Test file created with all required test cases
2. All 10 error codes (AG001-AG010) have dedicated tests
3. Enum value suggestions tested and verified accurate
4. Valid enum combinations tested via parametrize
5. File type filtering tested (validator skips non-agent files)
6. Coverage ≥90% for AgentEnumValidator class
7. All tests pass

#### Verification Steps

1. Run tests: `pytest plugins/plugin-creator/tests/test_agent_enum_validator.py -v`
2. Check coverage: `pytest --cov=plugin_validator --cov-report=term-missing`
3. Verify AG001 suggests valid model values
4. Verify validator doesn't run on skill files
5. Run `mypy tests/test_agent_enum_validator.py --strict`

**Can Parallelize With**: Task 15 (HookConfig tests), Task 16 (MCPConfig tests), Task 17 (LSPConfig tests)
**Reason**: Independent test files
**Handoff**: Provide test file, coverage report, suggestion verification

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.