The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using bun to interact with specialized portal-search tools located in the .agents/skills/ directory.
[DYNAMIC_EXECUTION]: The skill implements a plugin-style architecture where it discovers other skills at runtime, reads their SKILL.md files, and executes commands defined within them. This creates a dependency on the integrity and safety of other installed portal skills discovered in the local environment.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests large amounts of untrusted data from external websites and job portals.
Ingestion points: Job listings and descriptions are fetched from external URLs via WebSearch, WebFetch, and external CLI tools (found in SKILL.md Step 1 and Step 2).
Boundary markers: No specific boundary markers or instruction-ignoring delimiters are used when processing job descriptions for fit assessment.
Capability inventory: The agent has access to file writing tools (Write, Edit) and restricted shell execution (Bash) as defined in the SKILL.md frontmatter.
Sanitization: While the skill parses structured data, it does not explicitly sanitize the natural language content of job descriptions before fit evaluation or presentation.

scrape