Deep Research
v1.0.0Conduct exhaustive multi-source investigation with methodology tracking, source evaluation, and iterative depth.
⭐ 9· 8.3k·60 current·63 all-time
byIván@ivangdavila
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description ('Deep Research') matches the provided instructions and supplemental files (methodology, sources, output formats). The skill is instruction-only and does not require unrelated binaries, env vars, or config paths.
Instruction Scope
The SKILL.md prescribes broad research behaviors (multiple search engines, primary sources, forum/social mining, tracking queries and sources). That scope is appropriate for a deep-research skill, but it gives the agent wide latitude on where to look (including social/forums). The instructions do not direct reading local files, exfiltrating secrets, or contacting unknown endpoints; however, chasing social/forum content may implicitly require accessing rate-limited or authenticated APIs if the agent attempts private/archival content.
Install Mechanism
No install spec or code present (instruction-only). This minimizes filesystem/execution risk — nothing is downloaded or installed by the skill itself.
Credentials
The skill declares no environment variables, credentials, or config paths. That is proportionate to an instruction-only research methodology. Note: some sources the methodology recommends (private social media, paywalled databases) may require credentials in practice, but the skill does not request them.
Persistence & Privilege
always:false and no self-modifying install behavior. The skill does not request permanent presence or elevated privileges beyond normal autonomous invocation.
Assessment
This skill is essentially a structured checklist and templates for doing in-depth research — it's coherent and low-risk. Before using it, explicitly confirm the research scope, depth, and time budget so the agent doesn't run open‑ended searches. If you need the agent to access paywalled databases, private accounts, or archived social-media content, supply only the minimal credentials you trust and be aware of terms-of-service and privacy rules. Consider whether you want to allow autonomous invocation for long-running/exhaustive tasks; otherwise invoke it on demand and review the agent's sourced links and methodology logs before sharing any sensitive outcomes.Like a lobster shell, security has layers — review code before you run it.
latestvk972nars0321ft9w0yz1b0mp61810p3j
License
MIT-0
Free to use, modify, and redistribute. No attribution required.