The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted content from an external issue tracker (issue bodies and comments), creating a surface for indirect prompt injection attacks. Attackers could place malicious instructions in comments to manipulate the agent's logic during the triage process.
Ingestion points: Found in SKILL.md under the "Triage a specific issue" section which reads "full issue (body, comments, labels, reporter, dates)" from the project's external tracker.
Boundary markers: There are no specified delimiters or instructions to treat issue content as untrusted data within the prompt context.
Capability inventory: The skill explicitly authorizes the agent to execute shell commands and tests in SKILL.md (Step 3: Reproduce).
Sanitization: No logic is provided to sanitize or validate the content of the issues before the agent acts upon them.
[COMMAND_EXECUTION]: The skill instructions specifically direct the agent to execute shell commands and tests for bug reproduction (SKILL.md). This provides an execution vector for any instructions injected via the issue tracker data.

triage