The Agent Skills Directory

[SAFE]: The skill provides a structured reasoning framework for knowledge synthesis and factual checks. It does not contain any executable code, shell commands, or network operations.
[NO_CODE]: No scripts, binaries, or configuration files that trigger code execution are present. The agent acts solely as a reasoning assistant following the provided markdown instructions.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to ingest and process external sources and claims (an untrusted data surface), the instructions explicitly mandate a critical analysis approach. It requires the agent to evaluate authority, freshness, and contradictions, which reduces the risk of the agent following malicious instructions embedded in the audited text. Furthermore, the skill lacks any dangerous capabilities (like file writing or network requests) that could be exploited via injection.

knowledge-review