The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill features a vulnerability surface where the agent is instructed to ingest and process untrusted external data to perform actions like generating documentation, updating tracking systems, or performing security audits.
Ingestion points: Git diffs and pull request evidence (workflows/document/SKILL.md), Kubernetes API audit logs (workflows/analyzing-kubernetes-audit-logs/SKILL.md), and external vulnerability scan results (workflows/triaging-vulnerabilities-with-ssvc-framework/SKILL.md).
Boundary markers: The workflows include instructions to ground facts in actual evidence (workflows/document/SKILL.md) and perform evidence-based reviews (workflows/review/SKILL.md) to mitigate hallucinations or adversarial content.
Capability inventory: The skill can perform file writes for scaffolding, execute shell commands for testing/scanning (workflows/thermo-fix/SKILL.md), and interact with external APIs via MCP (workflows/linear/SKILL.md).
Sanitization: The procedures emphasize technical validation, such as verifying tool schemas before MCP usage (workflows/mcp-linear-planner/SKILL.md) and normalizing audit log fields (workflows/analyzing-kubernetes-audit-logs/SKILL.md).
[EXTERNAL_DOWNLOADS]: The skill documentation and templates recommend the installation and use of various legitimate third-party packages and tools.
Evidence: Recommends installing remotion and related Node.js packages for video generation (workflows/remotion-video-motion/SKILL.md). It also references the installation of security tools like Trivy, Kubesec, and Semgrep (workflows/implementing-devsecops-security-scanning/SKILL.md).
Installation: The README instructs users to add the skill using npx skills add devton/agentic-workflow-blueprint, which is a standard platform command for the author's own repository.
[COMMAND_EXECUTION]: The workflows instruct the agent to execute a wide array of CLI tools to perform project tasks.
Evidence: Execution of git commands for logging and committing (workflows/document/SKILL.md, workflows/thermo-fix/SKILL.md), infrastructure plan/apply commands (workflows/iac/SKILL.md), and security scanners (workflows/scanning-docker-images-with-trivy/SKILL.md). These are within the expected functional scope of a developer-oriented blueprint.

workflow-blueprint