The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): No instructions found that attempt to bypass safety guidelines, override system prompts, or extract internal agent instructions.
[DATA_EXFILTRATION] (SAFE): No sensitive file access (e.g., SSH keys, credentials) or unauthorized network operations were identified. Code snippets use placeholders like 'sendToMonitoring' for illustrative purposes.
[OBFUSCATION] (SAFE): No Base64, zero-width characters, homoglyphs, or other obfuscation techniques were detected in the text or code snippets.
[REMOTE_CODE_EXECUTION] (SAFE): No patterns for downloading and executing remote scripts (e.g., curl|bash) or unverified dependencies were found.
[COMMAND_EXECUTION] (SAFE): The skill does not contain executable command-line instructions or subprocess calls that could compromise the host system.
[PRIVILEGE_ESCALATION] (SAFE): No use of sudo, chmod, or other privilege-altering commands was identified.
[PERSISTENCE] (SAFE): No attempts to modify shell profiles, cron jobs, or startup services were found.
[METADATA_POISONING] (SAFE): The YAML frontmatter (name, description) is descriptive and contains no deceptive or malicious content.
[INDIRECT_PROMPT_INJECTION] (SAFE):
Ingestion points: The skill describes processing browser error events in 'SKILL.md', which are technically untrusted data sources.
Boundary markers: Not applicable to instructional code snippets.
Capability inventory: No active capabilities such as subprocess calls or file-writing are present in the skill.
Sanitization: Not applicable as the skill does not execute the provided examples.
[DYNAMIC_EXECUTION] (SAFE): No use of eval(), exec(), or runtime compilation techniques was detected.

intermittent-issue-debugging