awesome-claude
A curated list of awesome things related to Anthropic Claude
0Tools
9Findings
1.2kStars
—Downloads
Mar 22, 2026Last Scanned
Findings9
3critical
5high
0medium
1low
0informational
criticalC1Command InjectionMCP03-command-injectionAML.T0054
Pattern "`[^`]+`" matched in source_code: "`claude-opus-4-6`" (at position 2723)
Replace exec()/execSync() with execFile() and pass arguments as an array, never as a string. Validate all inputs against an allowlist before use in any shell context. For subprocess.run, always pass a list and shell=False.
criticalL1GitHub Actions Tag PoisoningMCP10-supply-chainAML.T0017
Pattern "(?:curl|wget|fetch)\s+.*\|\s*(?:bash|sh|node|python)" matched in source_code: "curl -fsSL https://claude.ai/install.sh | bash" (at position 8069)
Pin all GitHub Actions to full commit SHAs (e.g., uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29) instead of mutable version tags (@v3). Use tools like StepSecurity Harden-Runner or Pinact to auto-pin. Audit all workflow files for pipe-to-shell patterns. CVE-2025-30066 and CVE-2026-31976 demonstrated this enables secret exfiltration and RCE via CI/CD.
criticalQ13MCP Bridge Package Supply Chain AttackMCP10-supply-chainAML.T0054
Pattern "(?:mcp|fastmcp|langchain-mcp|llama-index-mcp)(?:>=|~=|==)?(?!\d)" matched in source_code: "MCP" (at position 805)
MCP bridge packages (mcp-remote, mcp-proxy, @modelcontextprotocol/sdk, fastmcp) are high-value supply chain targets — CVE-2025-6514 (CVSS 9.6) in mcp-remote affected 437,000+ installs. Always pin exact versions (no ^ or ~ ranges). Use lockfiles (package-lock.json, pnpm-lock.yaml, uv.lock). Never run `npx mcp-remote` without version pinning. Verify package integrity with `npm audit` or `pip-audit` before deployment. Reference: CVE-2025-6514, OWASP ASI04.
highK15Multi-Agent Collusion PreconditionsMCP05-privilege-escalationAML.T0054
Pattern "(agent|delegate|orchestrat).*(?:invoke|call|execute|spawn)(?!.*(?:rate[_\s-]?limit|throttle|quota|max[_\s-]?concurrent|semaphore))" matched in source_code: "agentic coding tool (CLI), with VS Code/JetBrains IDE integrations, Desktop GUI (Cowork), and browser control. Understands entire codebases, plans and execute" (at position 7403)
Implement collusion-resistant multi-agent architecture: (1) Verify agent identity cryptographically before accepting commands, (2) Apply ACLs to shared write surfaces, (3) Rate-limit cross-agent invocations, (4) Audit all inter-agent communication with timestamps and agent IDs, (5) Baseline normal interaction patterns for anomaly detection. Required by MAESTRO L7 and CoSAI MCP-T9.
highK19Missing Runtime Sandbox EnforcementMCP07-insecure-configAML.T0054
Pattern "(run[_\s-]?as[_\s-]?root|EUID.*0|uid.*0)(?!.*(?:drop|setuid|seteuid|change.?user))" matched in source_code: "uide to Claude Code 2.0 and getting better at using coding agents](https://sankalp.bearblog.dev/my-experience-with-claude-code-20-and-how-to-get-better-at-using-coding-agents/#setup) - Part 2: A deep dive into the 2.0" (at position 15794)
Run MCP servers in sandboxed containers with: (1) No --privileged flag, (2) Minimal Linux capabilities (drop ALL, add only needed), (3) Read-only root filesystem, (4) Non-root user, (5) Seccomp/AppArmor profiles enabled, (6) No host mounts except data volumes. CoSAI warns containers alone are insufficient — add seccomp profiles. Required by CoSAI MCP-T8 and ISO 27001 A.8.22.
highQ14Concurrent MCP Server Race ConditionMCP07-insecure-configT1068
Pattern "(?:read|write|modify|delete).*(?:file|path|directory)(?!.*(?:lock|mutex|semaphore|flock|atomic))" matched in source_code: "readme) - Slash-commands, CLAUDE.md file" (at position 9279)
MCP servers sharing filesystem or database backends with other servers must implement proper concurrency controls. Use: (1) file locking (flock/lockfile) for filesystem operations, (2) database transactions for all read-modify-write sequences, (3) atomic file operations (O_EXCL, mkdtemp) instead of check-then-create, (4) lstat() to detect symlinks before following (CVE-2025-53109). Never assume exclusive access to shared resources — other MCP servers may be operating concurrently.
highC11ReDoS VulnerabilityMCP07-insecure-config
Pattern "\([^)]*\)[*+][*+]" matched in source_code: "(https://aws.amazon.com/bedrock/anthropic/)**" (at position 5299)
Never compile user-supplied strings as regexes. Use a safe regex library (re2, node-re2) for all user-controlled pattern matching — they run in linear time. Bound all input lengths before regex application. Use static analysis tools (safe-regex, vuln-regex-detector) to audit existing patterns.
highK13Unsanitized Tool OutputMCP02-tool-poisoningAML.T0054
Pattern "(?:query|execute|select|find).*(?:return|respond|result|rows|data)(?!.*(?:sanitize|escape|encode|map|filter|select|pick))" matched in source_code: "executes multi-step tasks, edits files, runs git/shell, creates PRs, and integrates brows" (at position 7554)
Sanitize all external data before including in tool responses. Implement output encoding that neutralizes prompt injection patterns. Truncate excessively long content. Validate structure before passing database results. Apply the principle: treat all external data as untrusted, even in tool outputs. Required by CoSAI MCP-T4.
lowF4MCP Spec Non-ComplianceMCP07-insecure-config
Server fails MCP spec compliance checks: required:server_name; required:server_version; required:protocol_version; recommended:tool_descriptions; recommended:parameter_descriptions
Follow the MCP specification for server metadata. Include server name, version, and protocol version. Provide descriptions for all tools and parameters.
Tools
No tools exposed by this server.
Security Category Deep Dive
Sub-Category Tree · Remediation Roadmap · Attack Stories · Compliance Overlay · ATLAS Techniques · Maturity Model
Prompt Injection
Prompt & context manipulation attacks
69
Maturity
14
Rules
5
Sub-Categories
1
Gaps
64%
Implemented
56
Tests
1
Stories
100%3 rules
Injection via tool descriptions and parameter fields
GAP-001Prompt Injection Coverage GapMissing detection coverage for emerging prompt injection attack variants not addressed by current rules
100%4 rules
Hidden instructions via external content and tool responses
100%2 rules
Context window saturation and prior-approval exploitation
100%3 rules
Payload hiding via invisible chars, base64, schema fields
100%2 rules
Injection via prompt templates and runtime tool output