The Agent Skills Directory

[SAFE]: The skill is composed entirely of markdown instructions and does not include any executable code, scripts, or binary dependencies.- [SAFE]: No evidence of base64 obfuscation, hidden characters, or malicious prompt injection patterns (such as 'ignore previous instructions') was found.- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that ingests external, untrusted data including campaign snapshots, monitoring results, and search-term reports. Ingestion points: Normalized snapshots and search term reports (SKILL.md). Boundary markers: None explicitly defined in the prompt text. Capability inventory: Ability to apply bid changes, budget reallocations, and keyword actions through a mutation gate (SKILL.md). Sanitization: Changes are applied only after being approved as mutations, verified against remote state, and audited, with a default posture of drafting rather than direct execution. (File: SKILL.md)

ads-optimize