The Agent Skills Directory

[PROMPT_INJECTION]: The instructions do not contain commands to override agent behavior, bypass safety filters, or extract system prompts. The language is purely instructional and aligned with the stated purpose of standards compliance.
[DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file paths (such as .ssh or .env) were identified in the skill instructions or the audit template.
[REMOTE_CODE_EXECUTION]: The skill does not perform any external downloads, package installations, or command executions. It functions entirely as a documentation and auditing guide.
[OBFUSCATION]: All content in SKILL.md and the audit template is provided in plain text. There is no evidence of Base64 encoding, zero-width characters, or hidden Unicode tags.
[INDIRECT_PROMPT_INJECTION]: The skill ingests information about a software project (e.g., system name, target sector) to populate a report template. While this is an ingestion surface, the skill lacks dangerous capabilities—such as shell execution, network access, or file-writing tools—that could be exploited by malicious data in the audited project description.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not use any dynamic command execution syntax (e.g., !command) to run shell commands at load time.

global-standards