Deep Research 🔬
v2.0.0Deep web research with multi-round search, cross-verification, and structured reports with citations. Enhances web_search and web_fetch into a full research...
⭐ 0· 91·0 current·0 all-time
byWen@wyatt88
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match what is implemented: the script produces web_search and web_fetch tool calls, deduplicates and tiers sources, and produces a report skeleton. There are no unrelated credentials, binaries, or external services requested that would be inconsistent with a research workflow.
Instruction Scope
SKILL.md instructs the agent to run the included Python script and to execute the platform's web_search and web_fetch tools, collect JSON results, and write reports. The script itself only manipulates search/fetch data and generates report scaffolding. Points to note: the Deep mode mentions spawning sub-agents (parallel search sessions) which increases agent autonomy, and the SKILL.md expects the agent to write temp files (e.g., /tmp/search-results.json) and final reports to a research/ directory. There are no instructions to read unrelated system files or secrets.
Install Mechanism
No install spec — instruction-only with one bundled Python script. This is low-risk compared to arbitrary downloads. The contained script is readable and does not attempt to fetch or execute additional remote code.
Credentials
The skill declares no required env vars or credentials. The script does optionally read non-secret environment variables (RESEARCH_DIR, SEARCH_COUNT, FETCH_MAX_CHARS) and defaults to a path under the user's home (~/.openclaw/workspace/research). These are not secrets but will affect where files are written; users should be aware of file creation in their home directory.
Persistence & Privilege
The skill is not always-enabled and does not request elevated platform privileges. It writes reports to a workspace directory and temp files per the instructions but does not modify other skills or system-wide agent settings. Deep mode's spawning of sub-agents increases activity but is within the documented workflow.
Assessment
This skill appears coherent and implements what it claims: a planner → search → fetch → report pipeline using the platform's web_search/web_fetch tools and a local Python helper. Before installing, consider: 1) It runs a bundled Python script locally—if you want extra caution, inspect the full script (scripts/research.py) yourself or run it in a sandbox. 2) It will write temporary files and final reports by default under ~/.openclaw/workspace/research (or wherever RESEARCH_DIR points); set RESEARCH_DIR if you prefer a different location. 3) Deep mode can spawn parallel sub-agents (more autonomous activity); if you have strict limits on agent autonomy, avoid Deep or confirm platform-session policies. 4) No credentials are requested or needed. If you plan to run this on a sensitive host, either review the code line-by-line or run in an isolated environment.Like a lobster shell, security has layers — review code before you run it.
latestvk97ec11x54p35fvw5nzgsxb7g5839tqp
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🔬 Clawdis