linkedin-post-research
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [DATA_EXFILTRATION]: Communicates with the Crustdata API to fetch research data.
- Endpoint:
https://api.crustdata.comcalled inscripts/search_posts.py. - [PROMPT_INJECTION]: The skill ingests LinkedIn post content which could potentially contain indirect instructions.
- Ingestion points: Post text and metadata from the Crustdata API are processed in
scripts/search_posts.py. - Boundary markers: No delimiters or warnings are used to separate untrusted post content from system instructions.
- Capability inventory: The skill can write output to the local file system and print results to the console.
- Sanitization: The script truncates post previews to 200 characters and removes newlines, reducing the risk and complexity of potential embedded instructions.
Audit Metadata