The Agent Skills Directory

[NO_CODE]: The skill consists entirely of Markdown instructions and reference files. There are no scripts, binaries, or automated tasks associated with the skill.
[COMMAND_EXECUTION]: No shell commands, subprocesses, or system-level operations were detected in the instructions or reference materials.
[EXTERNAL_DOWNLOADS]: The skill does not perform any network operations or external downloads. All reference material is stored locally within the skill's reference directory.
[DATA_EXFILTRATION]: There is no evidence of data collection or exfiltration. The skill operates locally on user-provided text for the purpose of coaching and drafting.
[CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or sensitive credential paths were found in any of the analyzed files.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided content (drafts and descriptions) to provide coaching, it lacks any executable capabilities or access to sensitive tools that could be exploited via malicious input.
Ingestion points: The agent reads user-provided communication goals and drafts via standard interaction.
Boundary markers: None explicitly defined in instructions, though the agent is instructed to focus on storytelling diagnostic steps.
Capability inventory: No tool use, file-writing, or network operations are included in the skill definition.
Sanitization: Standard model-level sanitization applies; no custom filtering is implemented.

storyteller-guidance