tentacle-post2wechat
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
main.tsusesspawnSyncto execute a local image compression tool at.agents/skills/baoyu-compress-image/scripts/main.ts. - [EXTERNAL_DOWNLOADS]: The skill communicates with
https://api.tentacle.proto upload images and save drafts to WeChat. This domain belongs to the skill vendor and the communication is consistent with the stated purpose of the skill. - [PROMPT_INJECTION]: The skill identifies as having a surface for indirect prompt injection due to the way it processes external data.
- Ingestion points: The script reads and processes the full content of a user-provided HTML file (passed via CLI argument
args.input). - Boundary markers: There are no explicit boundary markers or instructions to the model to ignore embedded commands within the input HTML.
- Capability inventory: The skill can execute local scripts (
spawnSync) and make network requests (fetch) to the vendor's API. - Sanitization: The HTML content is parsed using regular expressions for title and summary extraction but is not sanitized against embedded malicious instructions before being used in the workflow.
Audit Metadata