The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a custom CLI tool named opencontext and standard git commands to manage context checkpoints, branching, and merging. It also references operations such as pip install and ls for environment management and verification.
[EXTERNAL_DOWNLOADS]: The documentation suggests the installation of the opencontext package from a standard package registry using pip. It also references a JavaScript plugin file (opencontext-reminder.js) for session monitoring.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of retrieving and merging historical context. If malicious instructions are captured in a commit during a session, they could be re-injected when the agent later uses opencontext context or opencontext merge.
Ingestion points: Data enters the agent's context through history retrieval commands like opencontext context and integration commands like opencontext merge.
Boundary markers: The instructions do not define specific delimiters or warnings to treat historical context as untrusted data.
Capability inventory: The skill facilitates file system writes (to the .GCC/ directory), execution of Git commands, and execution of the opencontext CLI.
Sanitization: No evidence of sanitization or validation of the historical data is present within the skill's instructions.

opencontext