The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the boss-agent-cli package and the patchright browser automation tool to function. These are identified as vendor-provided resources or standard industry tools for the described task.- [COMMAND_EXECUTION]: The skill operates by executing the boss command-line utility for search and interaction tasks. It also includes instructions for the agent to move session files and request permission elevation if restricted by sandbox environments.- [CREDENTIALS_UNSAFE]: The tool manages sensitive job platform sessions and AI API keys. These are stored locally in the /tmp/boss-agent-codex or ~/.boss-agent directories using standard local configuration patterns.- [PROMPT_INJECTION]: The skill contains a vulnerability surface for indirect prompt injection as it processes untrusted data from the recruitment platform.
Ingestion points: External content is ingested via boss detail, boss ai analyze-jd, and boss chatmsg.
Boundary markers: No explicit delimiters are used in the prompt instructions to separate untrusted platform content from the agent's core instructions.
Capability inventory: The agent possesses capabilities to write to the platform (boss greet, boss ai reply) and submit applications (boss apply).
Sanitization: No sanitization or validation of job descriptions or chat content is specified before the data is processed by the AI.

boss-agent-cli