The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the codebase, such as filenames, code snippets, and git logs, and provides this content to subagents for analysis. This ingestion occurs without sanitization or explicit boundary markers, potentially allowing malicious content within the analyzed codebase to influence agent behavior.\n
Ingestion points: Output from scripts/recon.sh and codebase files read by subagents in Phase 2.\n
Boundary markers: Absent.\n
Capability inventory: Subagents are authorized to execute shell commands (ls, find, grep, git, etc.) on the local project to gather evidence.\n
Sanitization: Absent.\n- [COMMAND_EXECUTION]: The skill executes a local shell script (scripts/recon.sh) and uses subagents to run a variety of command-line tools (find, grep, wc, cat, git) to extract metrics and evidence from the filesystem.\n- [EXTERNAL_DOWNLOADS]: The skill includes a recommendation for the user to install the btar package via npm from a third-party GitHub repository (https://github.com/jaredmcfarland/btar), which is not an officially trusted organization or well-known service.

codebase-readiness