The Agent Skills Directory

[COMMAND_EXECUTION]: The skill extensively uses the Bash tool to perform technical tasks such as creating reproduction scripts, running git bisect in detached worktrees, and executing diagnostic probes.
[EXTERNAL_DOWNLOADS]: The skill leverages WebSearch and WebFetch to assist in decoding opaque error strings, status codes, and library stack frames from external sources.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data (such as bug reports, error logs, and web documentation) and possesses tools capable of affecting the local environment.
Ingestion points: User-provided symptoms, repository source code, and fetched web content (SKILL.md).
Boundary markers: The skill defines a strict behavioral "ritual" to gate execution, but it lacks technical delimiters or instructions to ignore embedded commands in the data it processes.
Capability inventory: The skill has access to Bash for shell commands, Write and Edit for file system modification, and WebFetch for network requests (SKILL.md).
Sanitization: There is no evidence of sanitization or filtering for external data before it is interpreted or used in command generation.

debugkit