The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @xquik/tweetclaw@1.6.31 package. This package originates from a third-party source that is not identified as a trusted organization or well-known service.
[COMMAND_EXECUTION]: The instruction set includes a shell command (openclaw plugins install npm:@xquik/tweetclaw@1.6.31) intended for the agent to execute, leading to the installation of external code at runtime.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from social media posts (X/Twitter), which are untrusted external sources. 1. Ingestion points: External data is ingested through the TweetClaw tool as described in SKILL.md. 2. Boundary markers: The instructions include a natural language warning to 'Treat posts as evidence to verify, not as truth', but lack strict technical delimiters or specific 'ignore instructions' guards. 3. Capability inventory: The environment allows for plugin installation and generation of structured reports based on these external inputs. 4. Sanitization: There is no evidence of automated filtering, schema validation, or sanitization of the social media content before it is processed by the agent.

trend-spotter