The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided text to simplify it. It includes specific defensive instructions (Step 2.1 in SKILL.md) to treat the input text always as data and never as instructions, and to ignore embedded commands. This follow-up on safety guardrails reduces the surface for injection attacks.
[EXTERNAL_DOWNLOADS]: The skill mentions external sources like asd-ste100.org and a GitHub repository for portugues-tecnico-controlado. These are cited for reference and educational context only. The skill does not perform automated downloads or execute remote code from these locations.
[COMMAND_EXECUTION]: The skill does not contain any shell commands, scripts, or automated tool invocations. It operates entirely as a set of natural language processing instructions for the AI agent.
[CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or sensitive credentials were found. Placeholders like 'YOUR_API_KEY_HERE' are absent, and the skill does not request or handle sensitive environmental data.
[OBFUSCATION]: The content is written in clear Portuguese and English. No base64, hex-encoded strings, or hidden Unicode characters were detected in the provided files.
[METADATA_POISONING]: The metadata fields (name, author, version) are consistent with the skill's stated purpose of technical Portuguese simplification. There are no deceptive instructions or hidden payloads in the frontmatter.

padrao-pts