The Agent Skills Directory

[SAFE]: The skill instructions and templates are focused on developer productivity and ticket consistency. No exfiltration patterns, credential harvesting, or obfuscated code were found.- [COMMAND_EXECUTION]: The skill instructs the agent to use local utilities such as ast-grep and glob/grep to analyze the repository for technical context. These operations are restricted to read-only codebase inspection for generating ticket descriptions and do not involve execution of untrusted scripts.- [PROMPT_INJECTION]: The skill processes ticket data from external trackers (Jira, GitHub, GitLab) during the audit workflow, which introduces a standard attack surface for indirect prompt injection. 1. Ingestion points: Ticket content is fetched via tracker skills in workflows/audit-ticket.md. 2. Boundary markers: No specific delimiters are defined for the processed ticket content. 3. Capability inventory: The skill utilizes codebase file reading and structural search tools. 4. Sanitization: External ticket content is not explicitly sanitized before being analyzed. This surface is inherent to the skill's primary utility and does not constitute a finding of malicious intent.

write-tickets