agentleFS
zebbernzebbern/pocmap/.claude/skills/pocmap-agent/SKILL.md

pocmap-agent

Skill7 starsChanged 4 months agoMITView on GitHub

Use the PocMap Python package for CVE exploit discovery, vulnerability research, and bug bounty hunting. Provides 22 MCP tools and 13 CLI commands for looking up CVEs, finding exploits/PoCs, discovering recent vulnerabilities, product-based CVE discovery, CPE/CVSS analysis, bug bounty report lookup, and practice lab environments. Trigger when the user mentions CVE lookup, exploit discovery, PoC finding, vulnerability assessment, bug bounty research, security analysis, CPE to CVE conversion, EPSS scoring, KEV catalog checking, recent CVE monitoring, product vulnerability discovery, or security report generation.

  • Reads credentials
  • Installs packages

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use PocMap to look up CVEs, find exploits/PoCs, discover recent vulnerabilities, map products/packages to CVEs, check KEV/EPSS, find bug …

    line 16
  2. Very specific:

    Silent negatives: always read sources / error category / recent filter_stats.poc_check before concluding "none".

    line 81
  3. Very specific:

    Package dual IDs: use canonical_cve + aliases on discover_package_cves rows.

    line 84
  4. Very specific:

    Every tool returns a dict (structuredContent), not a JSON string.

    line 94

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

PocMap Agent Skill

Use PocMap to look up CVEs, find exploits/PoCs, discover recent vulnerabilities, map products/packages to CVEs, check KEV/EPSS, find bug bounty reports, and locate practice labs.

Source of truth: src/pocmap/ wins over this skill. The public Python API is synchronous. Full MCP contracts live in references/mcp_tools.md (canonical agent consumption guide). CLI flags: references/cli_commands.md.

Quick Start

pip install "pocmap[server]"   # CLI + MCP (mcp SDK 2.x / MCPServer via [server])
# from a clone: pip install -e ".[server,dev]"

pocmap lookup CVE-2021-44228
python -m pocmap --help        # authoritative CLI list (13 commands)

# MCP server (src/pocmap/mcp/ → pocmap-mcp via mcp_server facade)
uvx --from 'pocmap[server]' pocmap-mcp
pocmap-mcp                     # after install
python -m pocmap.mcp_server
from pocmap.services import CVEService

with CVEService() as svc:
    info = svc.get_cve_info("CVE-2021-44228")
    print(info.cvss, info.epss, info.kev_status)  # epss is 0–100 in the Python model

Decision Guide

GoalMCP toolCLI
PoC / exploit repos onlyfind_github_pocs (first)(shown in lookup)
Full assessment of known CVE ID(s)generate_json_reportpocmap lookup / bulk
Look up one CVE (metadata)lookup_cvepocmap lookup CVE-…
Recent CVEs / monitoringfind_recent_exploitspocmap latest --since 24h
CVEs for a deployed productdiscover_product_cvespocmap discover "Product"
CVEs for a dependency / lockfilediscover_package_cves—
GitHub PoCs (+ labels/trust_score/sources)find_github_pocs(shown in lookup)
Metasploit (multi-module; path/type)find_metasploit_module (limit>1)—
ExploitDB / Nucleifind_exploitdb_entry / find_nuclei_template—
Prioritize known CVE IDsgenerate_json_report → read triage—
How it is exploited (ATT&CK)get_attack_techniques—
Verify PoC is realverify_github_pocs (needs POCMAP_ALLOW_FETCH_POC_SOURCE=1)—
KEV / EPSScheck_kev_status / get_epss_score—
Bug bounty / labsfind_bug_bounty_reports / find_practice_labs / find_vulhub_dockerbugbounty / labs
CVE ↔ CPEcve_to_cpe / cpe_to_cvecpes / cpe2cve
HTML reportgenerate_html_reportpocmap bulk
Playbooksget_*_playbook—

Key Constraints

  • CVE ID: CVE-YYYY-NNNN+ (^CVE-\d{4}-\d+$). Lowercase is normalized.
  • Bulk / report: max 100 CVEs per call.
  • EPSS scales (convert at the boundary):
    • Filter min_epss on CLI / find_recent_exploits: 0–100
    • MCP normalized CVE fields (epss_score on lookup_cve, reports, recent): 0.0–1.0
    • get_epss_score: 0.0–1.0
    • Python CVEInfo.epss: 0–100
  • --since: 1h, 24h, 7d, 30d. Severity: critical|high|medium|low.
  • Product vs package: discover_product_cves = deployed product (nginx, Confluence). discover_package_cves = dependency/SBOM (PyPI/npm/Maven/…) — only tool with fix versions.
  • Silent negatives: always read sources / error category / recent filter_stats.poc_check before concluding "none". Empty + rate_limited/error means unknown, not none. Empty ATT&CK list means unmapped, not harmless.
  • Package dual IDs: use canonical_cve + aliases on discover_package_cves rows.

Env vars for MCP. Clients launch the server with a filtered env. Put GITHUB_API_TOKEN / NVD_API_KEY / POCMAP_* in the client config env block — shell exports do not reach MCP. Settings: src/pocmap/config.py (POCMAP_HTTP_TIMEOUT, not POCMAP_REQUEST_TIMEOUT; GITHUB_API_TOKEN, not POCMAP_GITHUB_TOKEN).

Error Handling (MCP)

Every tool returns a dict (structuredContent), not a JSON string. Failures use an error envelope — check error first:

error, error_type, category, retryable, context
# category: not_found | rate_limited | offline | network_error |
#           invalid_input | permission_error | not_enabled | unknown

Retry only when retryable is true (≈3 attempts with backoff), then surface suggestion/hint if present.

Architecture (brief)

CLI / MCP → services/ → clients/ → models (pydantic). Services are sync context managers. Key classes: CVEService, ExploitService (find_exploits / find_exploits_with_status), ReportService, RecentService, ProductDiscoveryService, PackageService, LabService, BugBountyService.

References

  • references/mcp_tools.md — all 22 MCP tools, resources, prompts, return shapes
  • references/cli_commands.md — all 13 CLI commands with real flags
  • GitHub: https://github.com/zebbern/pocmap

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.