Skills/ClawHub/ardi-skill

ardi-skill@1.0.0

BLOCKED· install only

Publisher claims: "Ardi — Agent Ordinals. 21,000 words. The first on-chain dictionary, inscribed by AI agents on Base mainnet. Agent-only worknet — no human has ever minted an..." Observed in sandbox: sends data to release-assets.githubusercontent.com.

Registry ClawHubPublisher yangyangxing · publishing for 197 daysInstalls 787Last detonated 16 h ago

Coverage

Bundled scripts ran with no arguments; no agent task scenario for this bundle.
Tested
install scripts, with decoy credentials in place · second launch with the clock 45 days ahead · 8 of 9 bundled scripts run with no arguments (the rest were not run: per-bundle cap), with decoy credentials in the environment · every file read, file written and host contacted by what ran
Not tested
use of the skill by an agent

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (3 scenarios; bundled scripts ran with no arguments; no agent task scenario for this bundle) on 2026-11-19 and scored it 47/100 (Blocked). Do not install it: the sandbox observed behaviour consistent with credential theft, exfiltration or persistence. Notable findings: Privileged sudo command; Privileged sudo command; Direct script execution from GitHub raw URL. Summary: sends data to release-assets.githubusercontent.com.

Behavior trace

Honeypot run · 1 s · 3 scenarios
t+CategoryObserved action
0.0sprocessinstall script run in the sandbox
0.5snetworkCONNECT tcp://api.github.com:443
0.5snetworkCONNECT tcp://github.com:443
0.5snetworkCONNECT tcp://release-assets.githubusercontent.com:443 · not in the allow-list, refused
0.5sprocessexec /usr/bin/mkdir
0.5sprocessexec /usr/bin/uname
0.5sprocessexec /usr/bin/curl
0.5sprocessexec /usr/bin/mktemp
0.5sprocessexec /usr/bin/rm
0.5sprocessexited 1: Error: download failed from https://github.com/awp-worknet/ardi-skill/releases/download/v0.5.19/ardi-agent-linux-x86_64-musl
0.5sprocessinstall script run in the sandbox
0.6snetworkCONNECT tcp://api.github.com:443
0.6snetworkCONNECT tcp://github.com:443
0.6snetworkCONNECT tcp://release-assets.githubusercontent.com:443 · not in the allow-list, refused
0.6snetworkCONNECT tcp://raw.githubusercontent.com:443 · not in the allow-list, refused
0.6sprocessexec /usr/bin/mktemp
0.6sprocessexec /usr/bin/curl
0.6sprocessexited 22: curl: (22) The requested URL returned error: 403
0.6sprocessbundled script run in the sandbox with no arguments
0.6sprocessexited 1: FAIL: ardi-agent not on PATH (looked in /tmp/ks-detonate-AmLlEq/scripts/home/.local/bin:/app/apps/web/node_modules/.bin:/root/.cache/node/corepack/v1/pnpm/10.28.0/dist/node-gyp-bin:/app/node
0.6sprocessbundled script run in the sandbox with no arguments
0.6sprocessexec /usr/bin/dirname
0.6sprocessexited 64:
0.6sprocessbundled script run in the sandbox with no arguments
0.6sprocessexec /usr/bin/dirname
0.6sprocessexited 65: ⚠ ardi-agent not on PATH — install the skill first
0.6sprocessbundled script run in the sandbox with no arguments
0.6sprocessexec /usr/bin/sed
0.6sprocessexec /usr/bin/tr
0.6sprocessexited 127:
0.6sprocessbundled script run in the sandbox with no arguments
0.7sprocessbundled script run in the sandbox with no arguments
0.7sprocessexec /usr/bin/rm
0.7sprocessexited 127: ./tools/auto-mine/uninstall.sh: line 11: systemctl: command not found
0.7sprocessbundled script run in the sandbox with no arguments
0.7sprocessexited 1: ./tools/auto-mine/runtime/claude.sh: line 16: 1: usage: claude.sh <prompt-file>
0.7sprocessbundled script run in the sandbox with no arguments
0.7sprocessexited 1: ./tools/auto-mine/runtime/hermes.sh: line 15: 1: usage: hermes.sh <prompt-file>
0.7sprocess8 of 9 bundled scripts run with no arguments (the rest were not run: per-bundle cap)
0.7sprocessscript backend runs setup only
0.7sprocessscript backend runs setup only
1.0sprocesssecond launch, clock +45 days, CI unset
1.2snetworkCONNECT tcp://api.github.com:443
1.2snetworkCONNECT tcp://github.com:443
1.2snetworkCONNECT tcp://release-assets.githubusercontent.com:443 · not in the allow-list, refused
1.2sprocessexec /usr/bin/uname
1.2sprocessexec /usr/bin/curl
1.2sprocessexec /usr/bin/mktemp
1.2sprocessexec /usr/bin/rm
1.2sprocessexited 1: Error: download failed from https://github.com/awp-worknet/ardi-skill/releases/download/v0.5.19/ardi-agent-linux-x86_64-musl
1.3snetworkCONNECT tcp://api.github.com:443
1.3snetworkCONNECT tcp://github.com:443
1.3snetworkCONNECT tcp://release-assets.githubusercontent.com:443 · not in the allow-list, refused
1.3snetworkCONNECT tcp://raw.githubusercontent.com:443 · not in the allow-list, refused
1.3sprocessexec /usr/bin/mktemp
1.3sprocessexec /usr/bin/curl
1.3sprocessexited 22: curl: (22) The requested URL returned error: 403

Static findings

MEDIUM
File permission modification commandtools/auto-mine/install.sh:81 · File permission modification command — agent changes file ownership/permissionschmod 600 "$ENV_FILE"Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
HIGH
Privileged sudo commandinstall.sh:119 · Privileged sudo command — agent may execute with elevated privilegeselif command -v sudo >/dev/null 2>&1; thenFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
HIGH
Privileged sudo commandinstall.sh:20 · Privileged sudo command — agent may execute with elevated privileges# Default install dir: ~/.local/bin if writable (no sudo needed); fallFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
HIGH
Direct script execution from GitHub raw URLscripts/bootstrap.sh:11 · Direct script execution from GitHub raw URL — unverified code sourceURL="https://raw.githubusercontent.com/awp-worknet/ardi-skill/main/install.sh"Fix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
File permission modification commandtools/auto-mine/config.example.env:4 · File permission modification command — agent changes file ownership/permissions# chmod 600 ~/.ardi-agent/auto-mine.envFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
Process kill commandsrc/cmd/run_loop.rs:15 · Process kill command — agent can terminate processes// SIGINT / SIGTERM finishes the current tick (so we don't kill a childFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
HIGH
Privileged sudo commandinstall.sh:121 · Privileged sudo command — agent may execute with elevated privilegessudo mv "${TMPFILE}" "${INSTALL_DIR}/ardi-agent"Fix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
Remote installer piped to shellinstall.sh:4 · Remote installer piped to shell# curl -fsSL https://raw.githubusercontent.com/awp-worknet/ardi-skill/main/install.sh | shFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST07
MEDIUM
File type combined with upload actionsrc/rpc.rs:134 · File type combined with upload action — agent may exfiltrate local data filesmatch http().post(url).json(&req).send() {Fix: Remove the call, or declare the host in frontmatter and explain in the README what is sent and why.
AST07
HIGH
Direct script execution from GitHub raw URLinstall.sh:4 · Direct script execution from GitHub raw URL — unverified code source# curl -fsSL https://raw.githubusercontent.com/awp-worknet/ardi-skill/main/install.sh | shFix: Pin versions and sources; vendor anything fetched at install time.
AST06
MEDIUM
Process kill commandsrc/cmd/commit.rs:297 · Process kill command — agent can terminate processes// window between "tx mined" and "state saved" where a process killFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
MEDIUM
File permission modification commandinstall.sh:114 · File permission modification command — agent changes file ownership/permissionschmod +x "${TMPFILE}"Fix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05
MEDIUM
Remote installer piped to shellinstall.sh:23 · Remote installer piped to shell# the first time anyone ran the curl|sh one-liner — broke theFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST07
HIGH
Privileged sudo commandinstall.sh:22 · Privileged sudo command — agent may execute with elevated privileges# Pre-v0.5.9 default was /usr/local/bin which prompted for sudo on macOSFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05
CRITICAL
Pipe-to-shell patterninstall.sh:4 · Pipe-to-shell pattern — downloads and immediately executes remote code# curl -fsSL https://raw.githubusercontent.com/awp-worknet/ardi-skill/main/install.sh | shFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST06

Cost to the agent

Instructions loaded
≈8,025 tok
1 file · 50 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.

What would make this Verified

33 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Contacts a host you did not declare≈ +23 on the totalList every host you talk to in SKILL.md frontmatter (`network: [api.example.com]`), or stop contacting the ones you cannot justify.
  2. Pipe-to-shell pattern≈ +9 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
  3. Unclaimed publisher≈ +8 on the totalClaim yangyangxing: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  4. Privileged sudo command≈ +4 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
  5. Direct script execution from GitHub raw URL≈ +4 on the totalPin versions and sources; vendor anything fetched at install time.
  6. File permission modification command≈ +2 on the totalRequest the narrowest permission that works; replace wildcards with the exact paths or hosts.
Publish a new version to ClawHub, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim yangyangxingPropose a patch to the maintainer