The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation instructs the installation of a global Node.js package @aicontextlab/cli from the NPM registry to provide context management capabilities.
[COMMAND_EXECUTION]: The skill extensively uses the oc command-line tool via shell execution for initializing repositories, managing local databases, and building search indices.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and search document content, which creates a surface for indirect prompt injection if malicious instructions are stored in the documentation files.
Ingestion points: Documents and search results are read into the agent's context through oc search, oc doc ls, and oc context manifest (documented in SKILL.md).
Boundary markers: No explicit markers or "ignore instructions" delimiters are specified in the provided instructions for handling retrieved content.
Capability inventory: The agent is granted access to powerful tools including Bash, Read, Write, Grep, and Glob (configured in SKILL.md).
Sanitization: No specific sanitization or validation routines for ingested document data are described in the skill scripts or documentation.

opencontext