curriculum-optimizer
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill uses instructions to override the agent's standard behavior, such as 'Do NOT ask the user questions' and establishing a specific persona ('autonomous curriculum management analyst'). While these are common for specialized skills, they represent a modification of the underlying agent's default operational guidelines.
- [DATA_EXFILTRATION]: The skill contains logic to access a platform-specific hidden directory (
~/.claude/projects/) to record execution metadata. While used for telemetry purposes, this involves accessing the user's home directory outside of the immediate project scope, which constitutes a data exposure risk. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from the codebase being audited.
- Ingestion points: Reads
package manifests,core data structures, and general source code files during Phase 1 in SKILL.md. - Boundary markers: The instructions do not define clear delimiters or use 'ignore embedded instructions' warnings when processing the target codebase.
- Capability inventory: The skill has the capability to read files from the local filesystem and write reports to the
docs/directory, as well as write telemetry to~/.claude/projects/. - Sanitization: No explicit sanitization or validation of the ingested code content is performed before it is used to generate the analysis report.
Audit Metadata