The Agent Skills Directory

[DATA_EXFILTRATION]: The skill includes functionality to contribute 'execution traces' and 'contribution bundles' to a remote GitHub repository (Forsy-AI/agent-apprenticeship). These traces often capture system environment details, code snippets, and file contents which may contain sensitive data. The configuration supports an auto-share automatic mode that could lead to unintended data exposure if not closely monitored.
[INDIRECT_PROMPT_INJECTION]: The skill pulls 'learning signals' and 'experience packs' from a community ecosystem to influence agent workflows, creating a surface for indirect prompt injection.
Ingestion points: The apprentice ecosystem pull and apprentice learn create commands retrieve external data from a remote repository (SKILL.md).
Boundary markers: No specific boundary markers or 'ignore' instructions are documented for the ingested experience data.
Capability inventory: The agent can execute shell commands via the apprentice run command and interact with multiple LLM providers (SKILL.md).
Sanitization: The documentation does not mention sanitization or validation of the external content before it is used to influence the agent's task execution.
[EXTERNAL_DOWNLOADS]: The installation instructions direct users to download and install a global Node.js package (agent-apprenticeship) and utilize the GitHub CLI for repository interactions.
[COMMAND_EXECUTION]: The CLI provides a mechanism for custom agent configuration using command templates (e.g., my-agent run --workspace {workspace}), which involves the execution of shell commands based on user-defined strings.

agent-apprenticeship-ecosystem