The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes data from external, untrusted sources during its research phase.
Ingestion points: External data is retrieved through the use of WebSearch and the Context7 MCP documentation tool as specified in Step 3 of SKILL.md.
Boundary markers: The instructions do not define any specific delimiters or 'ignore instructions' warnings to prevent the agent from being influenced by malicious content embedded in the search results.
Capability inventory: The skill possesses no high-risk capabilities; no file writing, arbitrary command execution, or network exfiltration patterns were found in the skill metadata or instructions.
Sanitization: There are no specified methods for sanitizing, filtering, or validating the content retrieved from external sources before the model processes it.- [NO_CODE]: The skill is composed entirely of markdown instructions and priority configurations, with no accompanying scripts or executable files, which inherently limits its ability to perform unauthorized system operations.

oracle