The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze external data, such as project PRDs and competitive documentation (e.g., docs/prd/*), as well as results from WebSearch and WebFetch. This creates a surface where malicious instructions embedded in these external sources could influence the agent's behavior. The skill lacks explicit instructions for the agent to treat this content as untrusted data or to use boundary markers during interpolation.
Ingestion points: docs/prd/* (referenced in SKILL.md, Step 1), WebSearch, and WebFetch operations.
Boundary markers: Absent in the instructions for data analysis.
Capability inventory: Bash tool, Read, Grep, Glob, WebSearch, WebFetch.
Sanitization: Absent.
[COMMAND_EXECUTION]: The skill provides a bash script (scripts/check-availability.sh) and instructions to execute it with user-derived candidate names. While the script employs proper shell quoting for variables, the process of interpolating user-supplied strings into shell commands via the agent's Bash tool presents a surface for command injection if the agent does not properly sanitize inputs containing shell metacharacters before execution.
[EXTERNAL_DOWNLOADS]: The scripts/check-availability.sh script performs network probes to verify name availability on several well-known services, including GitHub, PyPI, npm, crates.io, RubyGems, WordPress, and Telegram. These operations are essential to the skill's primary function and target established, reputable infrastructure.

naming