Caveat verdict
deep-recall
deeprecall
Recursive memory recall library that reads local markdown files and calls OpenAI-compatible LLM endpoints; package_install and agent_memory capabilities are used entirely for the stated purpose of persistent agent memory.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Permission integrity
package_install
agent_memory
Findings (7)
Accesses OpenClaw config/secrets directly
SKILL.md · frontmatter · ~/.openclaw/openclaw.json
Possible hardcoded credential
provider_bridge.py · prose · downgraded · api_key = "ollama-local
References agent memory files
SKILL.md · code · MEMORY.md
Social engineering — falsely claims user authorization
provider_bridge.py · prose · downgraded · the user has configured
Python httpx request — network access
deep_recall.py · prose · downgraded · httpx.post(
Python os.environ.get — reads environment variable
deep_recall.py · prose · downgraded · os.environ.get(
Python urllib.request — network access
provider_bridge.py · prose · downgraded · urllib.request
Permissions & capabilities
Requires 18 environment variables. (18 sensitive: ANTHROPIC_API_KEY (optional), OPENAI_API_KEY (optional), GOOGLE_API_KEY (optional), OPENROUTER_API_KEY (optional), DEEPSEEK_API_KEY (optional), MISTRAL_API_KEY (optional), TOGETHER_API_KEY (optional), GROQ_API_KEY (optional), FIREWORKS_API_KEY (optional), COHERE_API_KEY (optional), PERPLEXITY_API_KEY (optional), SAMBANOVA_API_KEY (optional), CEREBRAS_API_KEY (optional), XAI_API_KEY (optional), MINIMAX_API_KEY (optional), ZHIPU_API_KEY (optional), MOONSHOT_API_KEY (optional), DASHSCOPE_API_KEY (optional)).
package_installagent_memory Is this flag fair?
Thanks — recorded.