registry / AI-agent capability abuse / agentic-awesome-skills / 18.7.0

agentic-awesome-skills@18.7.0

AAS Core: complete local skill discovery, agent-owned selection, stack validation, and planning, backed by 2,474+ agentic skills.

AI Security Review

scanned 25d ago · by lpm-firewall-ai

The package has an explicit AI-agent skill installation capability that writes to assistant control directories. It has no automatic npm lifecycle hook, so the evidence supports a non-blocking capability warning rather than malware.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

A user runs the installer command without dry-run mode.

Impact

Installed skills can extend the capabilities and instructions available to the selected AI assistant.

Mechanism

The command resolves selected skill entries and copies them to the selected assistant skill directory.

AI rationale

Evidence

tools/bin/install.js~/.cursor/skills~/.claude/skills~/.gemini/skills~/.codex/skills~/.agents/skills

Decision evidence

public snapshot

AI called this Suspicious at 90.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

The manifest exposes an installer as a user-invoked command.
The installer selects several AI assistant skill directories, including Cursor and Claude Code locations.
The installer copies selected skill entries into its chosen target after the command is invoked.

Evidence against

No npm preinstall, install, or postinstall hook is declared.
The observed mutation is gated by explicit execution of the package command and target-selection options.

Behavioral surface

Source

ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShell

Supply chain

HighEntropyStringsUrlStrings

ManifestNo manifest risk signals triggered.

scanned 221 file(s), 1.50 MB of source, external domains: amazon.com, api.apify.com, api.expo.dev, browserbase.com, console.apify.com, developers.notion.com, docs.anthropic.com, docs.stripe.com, fonts.googleapis.com, github.com, graph.facebook.com, huggingface.co, kit.svelte.dev, nextjs.org, platform.openai.com, prod.spline.design, registry.npmjs.org, turborepo.dev, vercel.com, web.dev, www.w3.org

Source & flagged code

13 flagged

skills/developer-signup-flow/SKILL.mdView file

165patternName = stripe_test_secret severity = high line = 165

High

High Secret

Package contains a high-severity secret pattern.

skills/developer-signup-flow/SKILL.mdView on unpkg · L165

165patternName = stripe_test_secret severity = high line = 165

High

Secret Pattern

Stripe test secret key in skills/developer-signup-flow/SKILL.md

skills/developer-signup-flow/SKILL.mdView on unpkg · L165

tools/bin/install.jsView file

2L3: const { spawnSync } = require("child_process"); L4: const path = require("path");

High

Child Process

Package source references child process execution.

tools/bin/install.jsView on unpkg · L2

skills/electron-drive-skill/scripts/drive.mjsView file

133console.log(`Building (${config.build})...`) L134: const built = spawnSync(config.build, { cwd: PROJECT_ROOT, stdio: 'inherit', shell: true }) L135: if (built.status !== 0) fail(`'${config.build}' failed.`)

High

Shell

Package source references shell execution.

skills/electron-drive-skill/scripts/drive.mjsView on unpkg · L133

7Cross-file remote execution chain: skills/electron-drive-skill/scripts/drive.mjs spawns skills/electron-drive-skill/scripts/daemon.mjs; helper contains network access plus dynamic code execution. L7: */ L8: import { spawn, spawnSync } from 'node:child_process' L9: import { existsSync, mkdirSync, openSync, readFileSync, rmSync, statSync } from 'node:fs' L10: import http from 'node:http' L11: import path from 'node:path' ... L73: try { L74: const parsed = JSON.parse(data) L75: if (parsed.ok) resolve(parsed.result) ... L83: request.on('error', reject) L84: request.end(body) L85: }) ... L111: if (!existsSync(resolved) || !statSync(resolved).isDirectory()) return resolved

High

Cross File Remote Execution Context

Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.

skills/electron-drive-skill/scripts/drive.mjsView on unpkg · L7

tools/bin/aas.jsView file

3L4: const { main } = require("../lib/aas-v1/cli/main"); L5:

Medium

Dynamic Require

Package source references dynamic require/import behavior.

tools/bin/aas.jsView on unpkg · L3

skills/apify-content-analytics/reference/scripts/run_actor.jsView file

98const apiActorId = actorId.replace('/', '~'); L99: const url = `https://api.apify.com/v2/acts/${apiActorId}/runs?token=${encodeURIComponent(token)}`; L100: ... L102: try { L103: data = JSON.parse(inputJson); L104: } catch (e) { ... L114: }, L115: body: JSON.stringify(data), L116: }); ... L321: // Check for APIFY_TOKEN L322: const token = process.env.APIFY_TOKEN; L323: if (!token) {

Critical

Hardcoded Runtime Data Exfiltration

Source sends credentials or rich application records to a package-controlled external receiver enabled by default.

skills/apify-content-analytics/reference/scripts/run_actor.jsView on unpkg · L98

skills/huggingface-tool-builder/references/baseline_hf_api.tsxView file

39L40: const token = process.env.HF_TOKEN; L41: const headers: Record<string, string> = token ... L44: L45: const url = `https://huggingface.co/api/models?limit=${limit}`; L46: ... L54: L55: const text = await res.text(); L56: process.stdout.write(text); L57: })();

High

Credential Exfiltration

Source combines credential-like environment material and outbound requests; review data flow before blocking.

skills/huggingface-tool-builder/references/baseline_hf_api.tsxView on unpkg · L39

skills/gemini-deep-research/scripts/research.pyView file

•path = skills/gemini-deep-research/scripts/research.py kind = build_helper sizeBytes = 24204 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

skills/gemini-deep-research/scripts/research.pyView on unpkg

skills/web-artifacts-builder/scripts/shadcn-components.tar.gzView file

•path = skills/web-artifacts-builder/scripts/shadcn-components.tar.gz kind = compressed_blob sizeBytes = 19967 magicHex = [redacted]

Medium

Ships Compressed Blob

Package ships compressed or archive-like blobs.

skills/web-artifacts-builder/scripts/shadcn-components.tar.gzView on unpkg

skills/theme-factory/theme-showcase.pdfView file

•path = skills/theme-factory/theme-showcase.pdf kind = high_entropy_blob sizeBytes = 124310 magicHex = [redacted]

High

Ships High Entropy Blob

Package ships high-entropy non-source blobs.

skills/theme-factory/theme-showcase.pdfView on unpkg

skills/fedora-hyprland-installer/tests/test-scripts.shView file

•path = skills/fedora-hyprland-installer/tests/test-scripts.sh kind = payload_in_excluded_dir sizeBytes = 1740 magicHex = [redacted]

High

Payload In Excluded Dir

Package hides binary, compressed, or executable-looking payloads in test/fixture/hidden paths.

skills/fedora-hyprland-installer/tests/test-scripts.shView on unpkg

node_modules/ajv/dist/runtime/parseJson.js#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 8

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

node_modules/ajv/dist/runtime/parseJson.js#virtual:normalized:round1View on unpkg

Source & flagged code

13 flagged · loading source

skills/developer-signup-flow/SKILL.mdView file

165patternName = stripe_test_secret severity = high line = 165

High

High Secret

Package contains a high-severity secret pattern.

skills/developer-signup-flow/SKILL.mdView on unpkg · L165

165patternName = stripe_test_secret severity = high line = 165

High

Secret Pattern

Stripe test secret key in skills/developer-signup-flow/SKILL.md

skills/developer-signup-flow/SKILL.mdView on unpkg · L165

tools/bin/install.jsView file

2L3: const { spawnSync } = require("child_process"); L4: const path = require("path");

High

Child Process

Package source references child process execution.

tools/bin/install.jsView on unpkg · L2

Findings

1 Critical9 High6 Medium4 Low

CriticalHardcoded Runtime Data Exfiltrationskills/apify-content-analytics/reference/scripts/run_actor.js

HighHigh Secretskills/developer-signup-flow/SKILL.md

HighChild Processtools/bin/install.js

HighShellskills/electron-drive-skill/scripts/drive.mjs

HighCredential Exfiltrationskills/huggingface-tool-builder/references/baseline_hf_api.tsx

HighCross File Remote Execution Contextskills/electron-drive-skill/scripts/drive.mjs

HighShips High Entropy Blobskills/theme-factory/theme-showcase.pdf

HighPayload In Excluded Dirskills/fedora-hyprland-installer/tests/test-scripts.sh

HighSemantic Analysis Limitednode_modules/ajv/dist/runtime/parseJson.js#virtual:normalized:round1

High

Affected versions and remediation

This report applies to agentic-awesome-skills@18.7.0.

See all versions of agentic-awesome-skills for other recorded verdicts.

Review the evidence and your use of agentic-awesome-skills@18.7.0 before allowing it. Restrict the permissions described in this report.
Choose an independently verified alternative or release. This report does not establish that other versions are safe.

Evidence last updated: 2026-09-27 19:48:40.789Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Runtime surface

Behavioral surface

Source & flagged code