The Agent Skills Directory

[SAFE]: The skill is purely instructional and focused on financial and organizational modeling. It does not contain any executable scripts, binary files, or hidden commands.
[DATA_EXPOSURE_AND_EXFILTRATION]: No unauthorized data access patterns were identified. The skill instructions explicitly state the agent must 'remain read-only' and produce drafts or reviews unless explicit authority is granted by the owner.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests external data (client briefs and operating records), it lacks exploitable capabilities to facilitate an attack.
Ingestion points: Processes 'Client, plan owner, or approved project files' and 'operating records' as defined in SKILL.md.
Boundary markers: The instructions include controls to 'separate known facts from assumptions' and 'remain read-only'.
Capability inventory: The skill is restricted to document analysis and generation; it does not invoke shell commands, network utilities, or dynamic execution tools.
Sanitization: The workflow requires labeling all assumptions and citing sources, providing a layer of human-verifiable provenance for generated claims.
[COMMAND_EXECUTION]: No shell commands or system-level operations were detected in the instructions or references.

saas-sales-org-design-and-capacity-planning