iflow-mcp-crowdsecurity-crowdsec-local-mcp@0.6.0.post1.dev2
CAUTIONPublisher claims: "An MCP exposing prompts and tools to help users write WAF rules, scenarios etc." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios) on 2026-11-02 and scored it 69/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Fake system delimiter; Fake system delimiter; Fake system delimiter. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 1 s · 2 scenarios · 0 deviationst+CategoryObserved action
0.1sprocessno install scripts in bundle
0.5sprocessdependency install exit 1
0.6sprocessexec /usr/bin/timeout
0.6sprocessexec .venv/bin/crowdsec-mcp
0.6sprocessexited 127: timeout: failed to run command '.venv/bin/crowdsec-mcp': No such file or directory
0.6sclaimedno tools/list reply (exit 127)
0.6sprocessscript backend runs setup only
Static findings
CRITICAL
Fake system delimitersrc/crowdsec_local_mcp/prompts/prompt-waf-examples.txt:28 · Fake system delimiter — attempts to inject new instruction sectionFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01CRITICAL
Fake system delimitersrc/crowdsec_local_mcp/prompts/prompt-waf-examples.txt:225 · Fake system delimiter — attempts to inject new instruction sectionFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01CRITICAL
Fake system delimitersrc/crowdsec_local_mcp/prompts/prompt-waf-examples.txt:124 · Fake system delimiter — attempts to inject new instruction sectionFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01MEDIUM
Destructive file system command detectedsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:66 · Destructive file system command detectedFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05CRITICAL
Fake system delimitersrc/crowdsec_local_mcp/prompts/prompt-waf.txt:285 · Fake system delimiter — attempts to inject new instruction sectionFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01MEDIUM
Pipe-to-shell patternsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:64 · Pipe-to-shell pattern — downloads and immediately executes remote codeFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST06MEDIUM
File permission modification commandsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:40 · File permission modification command — agent changes file ownership/permissionsFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05CRITICAL
Fake system delimitersrc/crowdsec_local_mcp/prompts/prompt-waf-examples.txt:339 · Fake system delimiter — attempts to inject new instruction sectionFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_core.py:125 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_core.py:169 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Remote installer piped to shellsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:64 · Remote installer piped to shellFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST07MEDIUM
Process kill commandsrc/crowdsec_local_mcp/compose/waf-test/crowdsec/init-bouncer.sh:9 · Process kill command — agent can terminate processesFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05HIGH
Authorization escalation languagesrc/crowdsec_local_mcp/prompts/prompt-waf-examples.txt:194 · Authorization escalation languageFix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.
AST01MEDIUM
File permission modification commandsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:41 · File permission modification command — agent changes file ownership/permissionsFix: Request the narrowest permission that works; replace wildcards with the exact paths or hosts.
AST05MEDIUM
npx invocation of non-well-known package.github/workflows/build-mcpb.yml:35 · npx invocation of non-well-known package — may execute untrusted codeFix: Pin versions and sources; vendor anything fetched at install time.
AST06MEDIUM
Remote installer piped to shellsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:23 · Remote installer piped to shellFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST07MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_scenarios.py:334 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Pipe-to-shell patternsrc/crowdsec_local_mcp/compose/waf-test/nginx/Dockerfile:23 · Pipe-to-shell pattern — downloads and immediately executes remote codeFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST06MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_waf.py:400 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/setup_cli.py:181 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_waf.py:242 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_waf.py:111 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_waf.py:1012 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05MEDIUM
Process kill commandsrc/crowdsec_local_mcp/mcp_scenarios.py:415 · Process kill command — agent can terminate processesFix: Build commands from fixed arguments; do not interpolate untrusted strings into a shell.
AST05MEDIUM
Python subprocess execution detectedsrc/crowdsec_local_mcp/mcp_waf.py:78 · Python subprocess execution detectedFix: Build commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
AST05What would make this Verified
11 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%- Unclaimed publisher≈ +8 on the totalClaim pypi: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
- Does not do what the description says≈ +5 on the totalIn the sandbox the skill never performed the task it claims (no call to the service it is for). Make sure the happy path runs without extra setup, or describe the real behavior.
- New publisher≈ +3 on the totalTime clears this on its own after 14 and 30 days. Claiming the handle (+50) is the fastest way past it.
- Destructive file system command detected≈ +2 on the totalBuild commands from fixed arguments; do not interpolate untrusted strings into a shell.
- Pipe-to-shell pattern≈ +2 on the totalBuild commands from fixed arguments, never from interpolated strings; prefer an API client over shelling out.
- 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 PyPI, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim pypi