graphicode-junior-engineer-ts-flow
Pass
Audited by Gen Agent Trust Hub on Apr 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands (
catandecho) to interact with the local filesystem. It readsREADME.yamlfiles and writes generatedindex.tsfiles based on user-supplied directory paths. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted YAML data to construct TypeScript source code.
- Ingestion points: The agent reads the
README.yamlfile from arbitrary directories specified by the user in GraphiCode projects. - Boundary markers: The instructions lack explicit boundary markers or directives for the agent to ignore potentially malicious instructions embedded within the YAML structure.
- Capability inventory: The skill possesses file-write capabilities (
echo) which are used to generate code based on the ingested data. - Sanitization: There are no documented sanitization or validation steps to ensure the YAML content does not contain malicious code patterns or instructions that could corrupt the generated TypeScript files.
Audit Metadata