The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple sources to perform bug diagnosis, creating a potential surface for indirect instructions to reach downstream agents.
Ingestion points: Processes free-text user reports, attached screenshot descriptions, and external GitLab issues fetched via MCP as described in SKILL.md.
Boundary markers: The skill lacks explicit delimiters or specific "ignore instructions" warnings for the ingested content when processing the input for the diagnosis.
Capability inventory: Performs workspace file reads, executes container inspection commands, and constructs payloads with the diagnosis to dispatch other skills like ns-coder or ns-execution-gitlab-issue as seen in SKILL.md and references/continue-handoff.md.
Sanitization: No explicit sanitization or validation of the ingested external content is mentioned.
[COMMAND_EXECUTION]: The skill uses shell commands for environment inspection.
Evidence: The instructions permit the use of docker ps to verify container status during the investigation flow as specified in SKILL.md and references/investigation-map.md.

ns-bug-report-diagnosis