The Agent Skills Directory

[SAFE]: The skill follows established patterns for research tasks and includes explicit instructions for critical source evaluation. Local filesystem access is appropriately restricted to the 'docs/' directory.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it retrieves content from untrusted external sources (WebSearch, WebFetch) into the agent's context.
Ingestion points: External content is fetched in the GitHub, Reddit, and StackOverflow agents as specified in references/agent-prompts.md.
Boundary markers: The skill does not utilize specific delimiters or instructions to prevent the agent from obeying commands potentially embedded in fetched web content.
Capability inventory: The skill uses Grep (restricted path), WebSearch, WebFetch, and Task tools.
Sanitization: A 'Critical Evaluation' step (Step 7) is implemented to assess source credibility and recency, which serves as a logical data filter but not a security-focused content sanitizer.

research-online