openspec-apply-change

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell commands using the openspec CLI (e.g., openspec list, openspec status, openspec instructions). These commands are used to retrieve JSON metadata and manage the state of code changes within a project repository.
  • [DATA_EXFILTRATION]: The skill reads project-related files (proposals, specs, design, tasks) to provide context for implementation. While it accesses these files, it lacks instructions to transmit this data to external or untrusted domains.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted data from project files. 1. Ingestion points: Project files such as proposals, specifications, and task lists identified by the CLI's contextFiles output. 2. Boundary markers: The skill does not explicitly define markers to separate external data from its own instructional context. 3. Capability inventory: The skill has the capability to modify source code, update task status in markdown files, and execute shell commands through the openspec binary. 4. Sanitization: There is no explicit requirement for the agent to sanitize or validate the content of the project files before acting on implementation instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 12:04 PM
Security Audit — agent-trust-hub — openspec-apply-change