Crypto Research Interactive Framework
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a Markdown-only crypto research workflow with disclosed local workspace files, web research, and optional external data integrations.
This skill is generally coherent for crypto research. Before installing, decide whether you want collaborative or autonomous mode, verify any optional GitHub/MCP setup, keep API keys out of shared files, and periodically review or delete saved workspace notes.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Research queries and fetched pages may go through external search or fetch providers.
The skill expects the agent to use web/network research tools, which is appropriate for current crypto research but should be visible to the user.
Network — WebSearch / WebFetch ... Search for and fetch current crypto market data, news, documentation, and public blockchain data.
Use collaborative mode for sensitive topics, avoid putting private information into research queries, and review sources before relying on results.
If optional MCP servers are configured, the relevant provider credentials may be used for data lookups and may consume quotas.
Optional MCP integrations may use user-provided API keys for crypto data services; this is disclosed and purpose-aligned.
User manually configures each server and provides their own API keys
Use scoped API keys where possible, store them only in local config, and do not commit MCP configuration containing secrets.
Local workspace files may retain research context, user insights, and intermediate conclusions after a session ends.
The framework intentionally writes compressed research findings and user insights to local scratch state for continuity.
After user confirms checkpoint ... append final agreed findings to `.scratch` ... protects against context window compaction, enables session resume.
Review and delete workspace or scratch files when no longer needed, especially after working with sensitive source materials.
Using the GitHub setup route could cause the agent to read and follow remote Markdown instructions that may not match the installed package.
The README documents an optional remote-repository setup path, which can introduce provenance risk if the remote content differs from the reviewed registry artifact.
Give your AI assistant this link: https://github.com/kudodefi/crif ... The AI will clone the repo and activate CRIF automatically.
Prefer the reviewed package, or verify the repository and commit before asking an agent to clone and activate it.