Research Queue
v1.0.1Structured background research queue for unresolved technical, product, algorithmic, mathematical, and workflow questions. Use when the user wants to capture...
⭐ 0· 68·1 current·1 all-time
byMozi Arasaka@mozi1924
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description align with the instructions: the skill reads and updates a local QUESTIONS.md, can run small local experiments (exec/read), and fetch web resources. No unrelated binaries, env vars, or installs are requested.
Instruction Scope
The SKILL.md explicitly allows local reads and bounded execs and recommends writing to QUESTIONS.md and memory/YYYY-MM-DD.md. This is coherent for a research workflow, but it does give the agent the ability to read local files and run commands — including any paths listed in question 'links' (the example references ~/.openclaw/openclaw.json). Users should be aware the agent can access local files and networked pages as part of investigations.
Install Mechanism
No install spec and no code files are present. Instruction-only skills have low install risk because nothing new is written to disk by the skill package itself.
Credentials
The skill declares no environment variables, credentials, or config paths. However, its allowed-tools policy (read/exec/browser/web_fetch) enables reading arbitrary local files if a question links to them; the example includes ~/.openclaw/openclaw.json. Though plausible for debugging research, this could expose agent/config secrets if referenced in the queue.
Persistence & Privilege
always: false and no changes to other skills or global configuration are requested. The skill recommends using OpenClaw cron for autonomous runs and even prefers isolated runs, which limits cross-session pollution. Autonomous invocation is allowed (platform default) but not elevated here.
Scan Findings in Context
[no_regex_findings] expected: The regex scanner found nothing — expected because the skill is instruction-only with no code files to analyze.
Assessment
This skill appears coherent and appropriate for managing a research queue, but it permits the agent to read local files and run short commands as part of investigations. Before installing or enabling autonomous cron runs: (1) review QUESTIONS.md and any queued 'links' to ensure they don't reference sensitive config or credential files (e.g., ~/.openclaw/openclaw.json); (2) run the skill manually first so you can observe what files it reads and commands it runs; (3) prefer isolated cron runs (the skill recommends this) rather than embedding background work in your main session; and (4) avoid storing secrets or credentials in the queue or memory notes the agent will write. If you need stricter limits, restrict allowedTools (deny exec/read) or require manual invocation only.Like a lobster shell, security has layers — review code before you run it.
latestvk975qmwjdx4q2g6w3z7v5x86wn84a0kk
License
MIT-0
Free to use, modify, and redistribute. No attribution required.