agentleFS
Sign inStart free
ITch8ITch8/customs-data-skills/AGENTS.md

customs-data-skills

AGENTS.md9 starsChanged 24 days agoMITView on GitHub

Portable instructions for Codex, OpenClaw, Cursor, Claude Code, and other Agent Skills runtimes. Canonical skill files: skills/customs-lead-sop/, skills/customs-pitfalls/. The repository also contains other foreign-trade skills under skills/. This manual governs only the two customs skills. You help export sales teams use public customs / trade data to: You do not auto-send email, scrape behind logins, invent buyers from Comtrade, or invent buyer certainty from a single BOL row. No proprietary API key is required for the methodology. Paid trade-data subscriptions…

The file

We split this AGENTS.md into the rules it gives an agent

6 rules. About 1 is common, 2 are distinctive, and 3 are very specific to this project.

  • ≈1

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Avoid common false positives before outreach

    line 14
  2. Common:

    No proprietary API key is required for the methodology.

    line 34
  3. Very specific:

    Never mention customs/BOL/data platforms in customer-facing copy

    line 52
  4. Distinctive:

    No record ≠ no import

    line 56
  5. Very specific:

    Every buyer row must include: source_tools or source=user_export / secondary_lead, last_shipment_window (if any), evidence keywords/HS, …

    line 64
  6. Very specific:

    Use install.ps1 / install.sh in this package, or copy skills/* into the target root.

    line 77

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
AGENTS.md

Customs Data Skills — Agent Operating Manual

Portable instructions for Codex, OpenClaw, Cursor, Claude Code, and other Agent Skills runtimes.
Canonical skill files: skills/customs-lead-sop/, skills/customs-pitfalls/.
The repository also contains other foreign-trade skills under skills/. This manual governs only the two customs skills.

Role

You help export sales teams use public customs / trade data to:

  1. Prioritize markets (Type B stats — no company names)
  2. Find companies that have actually imported (Type A BOL where public; else secondary leads / user export)
  3. Spot supply-chain entry windows
  4. Avoid common false positives before outreach

You do not auto-send email, scrape behind logins, invent buyers from Comtrade, or invent buyer certainty from a single BOL row.

Required skills

SkillPathWhen
customs-lead-sopskills/customs-lead-sop/SKILL.mdFull workflow
customs-pitfallsskills/customs-pitfalls/SKILL.mdAlways with the SOP; also when reviewing lists/emails

Tool mapping (runtime-agnostic)

NeedPrefer
Open public pagesweb_fetch / WebFetch / browser
Discover pagesweb_search / WebSearch
Persist reportswrite files under ./hunt_result/ (or user output_dir)
Spreadsheetsread user CSV/Excel exports; local scripts / xlsx tools if present

No proprietary API key is required for the methodology. Paid trade-data subscriptions are optional upgrades the user may already have — prefer user export handoff over fighting anti-bot walls.

Default procedure

customs-lead-sop Steps 0–7
  (Step 2 = Type B stats only;
   Step 4 = Type A / user_export / secondary_lead)
        ↓
customs-pitfalls checklist
        ↓
Write customs-YY-MM-DD-{slug}.md (+ optional jsonl/csv)

Hard compliance (always on)

  • Public sources only (+ user-owned exports they already have rights to use)
  • Draft outreach only — never send
  • Never mention customs/BOL/data platforms in customer-facing copy
  • Translate customs fields → product / supply-chain insight in drafts
  • Consignee ≠ end-buyer without cross-check
  • Filter To Order / Bank / masked consignee; flag brokers & forwarders
  • No record ≠ no import
  • Statistical unit value ≠ deal price
  • Stats DB ≠ buyer list
  • Unverified HS → mark HS risk
  • Anti-bot / login wall → ask CSV/Excel; never fabricate rows

Output contract

Every buyer row must include: source_tools or source=user_export / secondary_lead, last_shipment_window (if any), evidence keywords/HS, score, tier, and flags when applicable:

possible_agent_or_3pl, possible_customs_broker, consignee_masked_or_to_order, possible_ams_confidentiality, possible_reexport_hub, hs_risk, evidence_thin, …

Install locations (reference)

RuntimeTypical skill root
Codex~/.codex/skills/ or .codex/skills/
OpenClaw~/.openclaw/skills/ or <workspace>/skills/
Cursor.cursor/skills/
Agents standard.agents/skills/ or ~/.agents/skills/

Use install.ps1 / install.sh in this package, or copy skills/* into the target root.

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.

Check your setup, free