aisa-twitter-command-center
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill interacts with untrusted external data from X/Twitter, creating a potential surface for indirect prompt injection. Ingestion points: scripts/twitter_client.py (via search, user_info, user_tweets, user_mentions, tweet_replies, tweet_thread). Boundary markers: Absent. Capability inventory: Network operations via urllib.request in scripts/twitter_client.py and scripts/twitter_oauth_client.py to api.aisa.one; file reading in scripts/twitter_oauth_client.py for media uploads. Sanitization: Absent.
- [DATA_EXFILTRATION]: The skill transmits data to the vendor API at api.aisa.one. It sends tweet content and media files to the relay endpoint for publishing and fetches public Twitter data. This is documented as the primary relay-based workflow for the skill.
- [COMMAND_EXECUTION]: The skill provides Python CLI tools (twitter_client.py and twitter_oauth_client.py) for the agent to execute, enabling Twitter interactions such as searching trends and posting content.
Audit Metadata