skills/atxp-dev/cli/atxp-memory/Gen Agent Trust Hub

atxp-memory

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill makes extensive use of npx to execute the atxp CLI tool for all operations including cloud synchronization (push, pull), status checks, and local indexing. The commands are directed at a user-provided directory path.
  • [EXTERNAL_DOWNLOADS]: The skill uses npx atxp@latest, which fetches and runs the latest version of the vendor's package from the NPM registry. It also downloads archived workspace data from the vendor's servers (atxp.ai) during the pull operation.
  • [REMOTE_CODE_EXECUTION]: The execution of npx atxp@latest constitutes the execution of remote code retrieved from a public registry at runtime. This is the primary mechanism for the skill's functionality.
  • [DATA_EXFILTRATION]: The push command is designed to upload the content of all markdown (.md) files within a specified directory to the vendor's cloud servers. While the skill documentation states only markdown files are sent, these files may contain sensitive user information or agent memories.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because the pull command retrieves content from a remote server and writes it to the local workspace. If the remote backup is compromised or contains malicious markdown, those instructions could be processed by the agent in future sessions.
  • Ingestion points: The pull command downloads data into the local filesystem at the specified --path (SKILL.md).
  • Boundary markers: None identified in the skill instructions to distinguish between trusted local data and pulled remote data.
  • Capability inventory: The skill has the capability to write to the filesystem and execute shell commands via npx.
  • Sanitization: There is no mention of sanitizing or validating the contents of the pulled markdown files before they are written to the workspace.
  • [DYNAMIC_EXECUTION]: The skill relies on npx to dynamically resolve and execute the atxp package. Additionally, it mentions an optional dependency, @zvec/zvec, which is an embedded vector database used for local indexing and search operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 04:35 AM