ClawHub

ClawForage Research Agent

v0.1.0

Deep domain research — entity extraction, cross-article connections, and structured domain reports from your knowledge base

⭐ 0· 99·0 current·0 all-time
byInspireHub.ai@dainash
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
ℹ
Purpose & Capability
The skill's name/description match its behavior: it reads memory/knowledge/*.md, extracts entities, builds cross-article connections, and writes reports to memory/research. One minor mismatch: metadata and README list jq as a required binary, but none of the included scripts actually call jq; bash and grep are used as expected.
✓
Instruction Scope
SKILL.md instructs only local operations on memory paths (reading knowledge articles, reading/writing memory/clawforage and memory/research). It explicitly constraints: read-only on knowledge articles and summaries-only rules. The instructions do reference placeholders ({baseDir}, {domain-slug}) which must be resolved by the runtime; otherwise commands will fail, but this is an implementation issue rather than a security concern.
✓
Install Mechanism
Instruction-only skill (no install spec). Included code are simple Bash scripts that operate locally and do not download or execute remote code. No risky download/extract steps are present.
✓
Credentials
No environment variables, credentials, or config paths are requested. The scripts operate on local memory paths only. This is proportionate to the stated research/reporting purpose.
✓
Persistence & Privilege
The skill is not always-enabled and does not request elevated agent-wide privileges. It writes report outputs and optional source whitelist files under memory/, which is consistent with its purpose and does not modify other skills or system-wide settings.
Assessment
This skill appears to do what it says: local entity extraction and report generation from harvested Markdown articles. Before installing: (1) ensure you run the Knowledge Harvester first so memory/knowledge/ contains files; (2) verify your OpenClaw runtime will substitute placeholders like {baseDir} and {domain-slug} when invoking the scripts; (3) note the README/metadata mention jq though the scripts don't use it—this is likely leftover and harmless but you can remove the jq requirement if you control the skill; (4) review the templates and scripts to confirm they meet your summary/attribution policies (the scripts do not enforce the 'max 10 words' rule—this is an agent-level behavioral constraint); (5) run the scripts in a sandbox or test environment first to confirm they behave as expected and only modify the intended memory/ directories.

Like a lobster shell, security has layers — review code before you run it.

latestvk97ed33j3nq68s599z6ymwf2yx8356mv

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

Binsjq, bash, grep

Comments