threads
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the Threads platform, including posts, comments, and search results, which could serve as a vector for indirect instructions.
- Ingestion points: The skill fetches data through commands like activity-feed, feed, search, and fetch-post-comments.
- Capability inventory: The agent has the ability to execute threads-cli commands, including the publish-post action which can create new content.
- Boundary markers: Instructions mandate using filtered structured results and discourage reconstructing omitted fields.
- Sanitization: The publish-post logic includes validation for file paths, basenames, and alt text length to prevent misuse.
- [COMMAND_EXECUTION]: The skill uses exec to invoke a specific companion binary, threads-cli. This is the intended execution model for the skill and is scoped to a predefined list of targets.
Audit Metadata