automating-api-testing
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/generate_test_suite.pyscript assembles executable shell scripts from input strings and useschmod 0o755to grant execution permissions. This allows the generation and potential execution of dynamically created command sequences. - [PROMPT_INJECTION]: The skill's workflow (Instructions 1 and 2 in
SKILL.md) involves reading and parsing external API specifications such as OpenAPI and GraphQL schemas. This represents an indirect prompt injection surface where a malicious specification could influence the agent's behavior. - Ingestion points: Parsing of OpenAPI and GraphQL specifications as defined in
SKILL.md. - Boundary markers: The instructions lack specific delimiters or warnings to treat specification content as untrusted data.
- Capability inventory: The skill possesses
Bash,Write, andEdittool permissions, which could be misused if the agent is misled by injected instructions. - Sanitization: No explicit sanitization or validation of input specification content is specified.
Audit Metadata