The Agent Skills Directory

[PROMPT_INJECTION]: Surface for Indirect Prompt Injection detected. The skill retrieves content from various Notion databases and incorporates it into reports and summaries.
Ingestion points: Multiple database queries in SKILL.md fetch task names, member names, and goal progress from the Notion workspace.
Boundary markers: The skill lacks explicit boundary markers (e.g., XML tags or delimiters) when interpolating retrieved data into the summary strings returned to the agent.
Capability inventory: The skill is configured with Read, Write, and Bash tools, which could potentially be abused if an agent follows instructions embedded in task or goal names.
Sanitization: There is no evidence of sanitization or escaping of the data fetched from Notion before it is rendered in the output.
[SAFE]: No hardcoded credentials, malicious network operations, or obfuscation techniques were found. All data processing remains within the expected Notion ecosystem.

10x-analytics