The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the OpenContext CLI package @aicontextlab/cli from the NPM registry. This downloads and installs third-party software required for the skill's functionality.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run various oc (OpenContext) commands for initialization (oc init), configuration (oc config), and database indexing (oc index build). These operations interact with the local file system, specifically creating and managing files within the ~/.opencontext directory.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection because it ingests and processes document content through its search and manifest features. If documents stored in the OpenContext database contain adversarial instructions, they could influence agent behavior when retrieved during a session.
Ingestion points: External data is ingested via oc search and oc_manifest commands from documents stored in the ~/.opencontext/contexts directory.
Boundary markers: The skill does not explicitly define delimiters or instructions to ignore embedded commands within the retrieved context.
Capability inventory: The skill possesses the capability to execute shell commands (Bash), read files (Read), and write to the file system (Write).
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the persistent storage before it is provided to the agent context.

opencontext