The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): No malicious override patterns, jailbreak attempts, or instructions to bypass safety filters were detected. The instructional language is benign and aligns with the stated purpose of leadership coaching.
[DATA_EXFILTRATION] (SAFE): There are no scripts or network operations present. The skill explicitly instructs the agent not to request secrets, credentials, or sensitive health data, and suggests using anonymized summaries for calendar audits.
[REMOTE_CODE_EXECUTION] (SAFE): No external dependencies, package managers, or remote script download patterns are used. The skill is entirely composed of Markdown files.
[COMMAND_EXECUTION] (SAFE): No shell commands, subprocess calls, or system-level operations are included in the skill workflow.
[INDIRECT_PROMPT_INJECTION] (SAFE): While the skill ingests untrusted user data (calendar text dumps and narratives), it lacks any high-privilege capabilities (like network access or arbitrary file writing) that could be exploited. The risk is limited to the local chat context.
[OBFUSCATION] (SAFE): All content is provided in clear-text Markdown with no encoded payloads, hidden characters, or homoglyphs detected.

energy-management