The Agent Skills Directory

[SAFE]: The skill follows established workflows for website information architecture. No security issues such as hardcoded secrets, obfuscated content, or unauthorized privilege escalation attempts were found in the instructions or configuration.
[NO_CODE]: The skill consists entirely of markdown and YAML files. It does not execute or download any scripts, binaries, or external dependencies, reducing the risk of supply chain or remote code execution attacks.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted project briefs. Evidence chain: (1) Ingestion points: Reads '00-project.md' and user-provided briefs in SKILL.md; (2) Boundary markers: Absent; (3) Capability inventory: Writing to local markdown files and interacting with external design tools; (4) Sanitization: Absent. This vulnerability surface is intrinsic to the skill's purpose as a document processing tool and is mitigated by the limited scope of its output capabilities.

webcraft-plan