Security audit
Security checks for vulnerabilities and agentic risk
This is a local OpenAPI changelog helper with no evidence of hidden execution, exfiltration, persistence, or destructive behavior.
Install only if you want an agent to read specified local OpenAPI or Swagger files and generate changelog text. Prefer explicit prompts like /api-changelog, and verify JSON, git comparison, and deprecation features before relying on them in CI.
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
## 🔴 Breaking Changes
### `DELETE /users/{id}` — REMOVED
The endpoint for deleting users has been removed.
**Migration:** Use `PATCH /users/{id}` with `{ "status": "deactivated" }` instead.
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
## 🔴 Breaking Changes
### `DELETE /users/{id}` — REMOVED
The endpoint for deleting users has been removed.
**Migration:** Use `PATCH /users/{id}` with `{ "status": "deactivated" }` instead.
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
API Changelog: openapi-v1.yaml → openapi-v2.yaml
🔴 4 BREAKING CHANGES:
- DELETE /users/{id} removed → use PATCH with status: deactivated
- GET /orders response: removed field order_total
- POST /checkout: new required param idempotency_key
- GET /products/{id}: price type changed number → string
The manifest explicitly says the skill outputs JSON for CI pipelines, and the usage examples include a --json mode. Yet the code shown stops at building ApiChange objects and includes only a Markdown report example; there is no serializer, CLI flag handling, or JSON emission implementation, creating a clear documentation-to-code contradiction.
The trigger list includes broad phrases like 'breaking changes' and 'what changed in the API', which can cause accidental invocation outside the intended context. In agent environments, over-broad triggers can route unrelated prompts into this skill, increasing the chance of unintended file handling, incorrect outputs, or misuse of local repository content.
The usage documentation advertises inputs like --vs HEAD~1, --vs main, and v2.0.0:openapi.yaml, implying the skill can resolve specs from git history or tagged revisions. However, the only implemented loader reads Path(path).read_text(...) from a filesystem path and there is no code to parse --vs, invoke git, or resolve revision-qualified spec paths, so the documented behavior contradicts the actual implementation shown.
The sample outputs include a deprecations section and counts, suggesting the skill detects deprecated endpoints or fields as part of its changelog generation. In the actual diff code, only endpoint, parameter, response-schema, and security changes are analyzed; there is no inspection of OpenAPI deprecated flags or sunset metadata, so the documentation overstates what the code does.