The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including local source code repositories and live WordPress posts.
Ingestion points: Local .go files are scanned and read during the DISCOVER and DISPATCH phases of the SAPCC review and audit workflows (SKILL.md). External WordPress posts are navigated to and evaluated using browser tools (SKILL.md, references/wordpress-phase-checks.md).
Boundary markers: The skill does not implement boundary markers or instructions to isolate untrusted data, increasing the risk that embedded instructions in the analyzed content could influence agent behavior.
Capability inventory: The skill utilizes powerful tools including Bash, Write, Edit, Agent, and Task which could be exploited through a successful injection (SKILL.md).
Sanitization: Content from analyzed files and web pages is processed without explicit sanitization or filtering.
[COMMAND_EXECUTION]: The skill executes local system commands to build, vet, and test code as part of its auditing and fixing workflows.
Evidence: Phase 4 (FIX) of the SAPCC Review workflow executes go build ./..., go vet ./..., make check, and go test ./... (SKILL.md). These commands are necessary for the skill's stated purpose of providing automated code reviews and fixes.
[DYNAMIC_EXECUTION]: The skill performs runtime compilation of Go code to verify fixes.
Evidence: The skill triggers go build and go test on the analyzed repository to verify fixes (SKILL.md). This is a standard part of the development and review workflow.

domain