The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute local scripts and testing commands within the repository environment.
Evidence: pnpm exec tsx scripts/cherry-regression-test/cli.ts run-phase and pnpm test:e2e:regression --list in SKILL.md.
Context: These commands are used to drive the testing framework developed by the vendor.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates testing of features (Chat, Agents, Knowledge Bases) that process external or untrusted data.
Ingestion points: The documentation mentions testing Chat, Agents, Skills, and knowledge bases.
Boundary markers: Not specified, though the skill emphasizes deterministic tests and platform-owned helpers.
Capability inventory: The process involves controlling an Electron instance and running TypeScript scripts.
Sanitization: Not explicitly mentioned in the skill instructions.
[CREDENTIALS_UNSAFE]: The skill defines environment variable names for authentication but provides safety rules to prevent their exposure.
Evidence: Reference to CHERRY_TEST_CUSTOM_PROVIDER_API_KEY and other sensitive variables in SKILL.md.
Context: The skill explicitly warns against printing credentials or including them in test artifacts like Playwright Traces.

cherry-regression-test