Caveat verdict
local-researcher
Reads local files AND makes external network calls
Privacy-first local research assistant using Ollama/LMStudio — all LLM inference runs locally, web searches use DuckDuckGo or optional API-keyed services; network access is for legitimate research queries and model setup, with no exfiltration of local secrets or files.
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.
What it does
These are capability combinations: each listed behavior occurs in the skill, but Caveat detects co-occurrence — it does not verify that one flows into another. Read the code to confirm a live chain.
Reads local files AND makes external network calls — the capabilities for data exfiltration co-occur (data-flow not verified)
LLM02 · LLM06 · ASI03
Accesses credentials AND makes external network calls — potential credential theft
LLM02 · ASI03
Permission integrity
network_out
file_read
credential_access
package_install
Findings (3)
Pipe-to-shell pattern (curl | sh) — supply chain attack vector
SKILL.md · code · curl -fsSL https://ollama.com/install.sh | sh
Pipe to sh — executes piped content as shell commands
SKILL.md · code · | sh
subprocess execution — runs system commands from Python
scripts/local-researcher-quickstart.py · prose · downgraded · subprocess.run(
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class A/D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
No declared permissions — minimal attack surface.
network_outpackage_installcredential_accessfile_read Thanks — recorded.