Dev Trace

Turn an email into an enriched, cited public profile—in one command.

A standalone agent skill from Built for Devs. Give it an email; it runs a deterministic waterfall of public sources—validate, surface, verify, GitHub, company enrichment, person enrichment, and a cited bio—and returns one structured profile with a confidence score and sources on every field.

• Works on day one with zero keys, and gets richer as you add them.
• Costs nothing by default. Paid providers are opt-in, so you spend only when a lead is worth it.
• Public data only. Below its confidence threshold it returns candidates, never invents.
• Zero dependencies (Node 18+). Runs the same way every time.

---

Quickstart

git clone https://github.com/built-for-devs/dev-trace-skill
cd dev-trace-skill

# Free layers only, no keys, no cost:
node scripts/dev-trace.mjs someone@example.com --pretty

That already validates the address and pulls Gravatar, domain, and GitHub identity. Add keys (below) to unlock company data, verification, person enrichment, and a cited bio.

Example

node scripts/dev-trace.mjs jordan@acme.dev --pretty

{
  "meta": { "email": "jordan@acme.dev", "depths_run": [0,1,2,3,4,5], "match_status": "verified", "sixtyfour_confidence": 9 },
  "profile": {
    "name":             { "value": "Jordan Lee",                        "confidence": 0.90, "sources": ["hunter","sixtyfour"] },
    "role":             { "value": "Head of Platform",                  "confidence": 0.90, "sources": ["hunter"] },
    "seniority":        { "value": "executive",                         "confidence": 0.85, "sources": ["hunter"] },
    "location":         { "value": "Austin, Texas, United States",      "confidence": 0.90, "sources": ["sixtyfour"] },
    "avatar":           { "value": "https://gravatar.com/avatar/...",   "confidence": 0.80, "sources": ["gravatar"] },
    "bio":              { "value": "Jordan Lee leads platform engineering at Acme ... [1][2]", "confidence": 0.85, "sources": ["tabstack-research"] },
    "company_name":     { "value": "Acme",                              "confidence": 0.90, "sources": ["tabstack","hunter"] },
    "company_industry": { "value": "Developer tools",                   "confidence": 0.85, "sources": ["tabstack"] },
    "company_domain":   { "value": "acme.dev",                          "confidence": 0.90, "sources": ["domain"] },
    "website":          { "value": "https://acme.dev",                  "confidence": 0.81, "sources": ["sixtyfour"] },
    "phone":            { "value": "+1 512 555 0143",                   "confidence": 0.90, "sources": ["sixtyfour"] },
    "social_linkedin":  { "value": "https://linkedin.com/in/jordanlee", "confidence": 0.90, "sources": ["hunter"] },
    "social_github":    { "value": "https://github.com/jordanlee",      "confidence": 0.90, "sources": ["sixtyfour"] }
    // ...and more, depending on which layers ran
  }
}

Every field carries its own confidence, the sources that produced it, and a conflict marker when providers disagree—so you always know how much to trust each value.

Fields it can populate

What you get back depends on which layers run, but the waterfall can fill a wide profile:

• Identity—name, role, seniority, location, avatar, bio
• Company—company_name, company_description, company_industry, company_size, company_location, company_domain
• Social—social_linkedin, social_twitter, social_github, any Gravatar-linked accounts, plus company_social_*
• Contact—email_found, phone, website
• GitHub—github_login, github_public_repos, github_followers

Fields only appear when a source actually returns them (public data only), and each is wrapped with value / confidence / sources / conflict.

What it costs

Dev Trace is built so you get a genuinely useful profile for free, and every paid provider is opt-in.

Tier	Provider	What you get
🟢 Free, no account	built-in	Email validation (syntax/MX/disposable), Gravatar + domain surface, GitHub identity
🟢 Free key, optional	GitHub token	Higher GitHub rate limits (GITHUB_TOKEN is free to generate)
🟢 Free account, goes far	Tabstack	Company/web enrichment and a cited, multi-source /research bio
🟡 Metered, free tier then paid	Hunter	Email verification + person enrichment (LinkedIn, role, seniority)
🔴 Premium, real $$$$	SixtyFour	Deepest person + LinkedIn dive. ~$4.80/pull (≈20 credits at medium, $0.24/credit), ~5/month on the entry plan. The break-glass layer, opt-in via --deep.

Run the whole free path and never spend a cent. The metered and premium layers are there for when you want the depth and choose to make the spend.

The waterfall

Depth	Layer	Source	Key	Runs
0	validate	syntax + MX + disposable	none	always, free
1	surface	Gravatar + domain scrape	none	always, free
1	verify	Hunter email verification	HUNTER_API_KEY	if key
2	profile	GitHub identity + basics	GITHUB_TOKEN optional	always, free
3	enriched	Tabstack web/company	TABSTACK_API_KEY	if key
4	deep	Hunter person enrichment	HUNTER_API_KEY	if key
5	deepest	SixtyFour people-intelligence	SIXTYFOUR_API_KEY	only with --deep
bio	synthesis	Tabstack /research (cited)	TABSTACK_API_KEY	when a name is resolved

Layers run in order and merge into one profile. A layer whose key is absent is skipped; a layer that errors is recorded in meta and never aborts the rest.

Get your keys

Every key is optional. Add only the ones you want, then drop them into .env.

Provider	Where to get a key	Unlocks	Tier
GitHub	github.com/settings/tokens	Higher GitHub rate limits	Free
Tabstack	tabstack.ai	Company enrichment + cited /research bio	Free account, generous
Hunter	hunter.io/api-keys	Email verification + person enrichment	Metered
SixtyFour	sixtyfour.ai	Deepest person + LinkedIn (--deep)	Premium

cp .env.example .env
# add the keys you have; the engine reads .env from the skill root or scripts/

.env is gitignored—keys are never committed.

Install as a skill

Dev Trace is a self-contained skill folder. To use it inside an agent, clone it into the directory that agent loads skills from:

git clone https://github.com/built-for-devs/dev-trace-skill ~/.claude/skills/dev-trace

For Claude Code that's ~/.claude/skills/ (global) or a project's .claude/skills/. The agent reads SKILL.md and invokes the engine for you. Not using a skills-aware agent? Run the engine directly, as in Quickstart.

Or install it through the Claude Code plugin marketplace:

/plugin marketplace add built-for-devs/dev-trace-skill
/plugin install dev-trace@built-for-devs

Using with other agents

The engine is a plain Node CLI, so any agent or script that can run a shell command can use Dev Trace—Codex, Cursor, Gemini CLI, OpenClaw, or your own automation.

• Claude Code reads SKILL.md.
• Codex and any AGENTS.md-aware tool read AGENTS.md.
• Gemini CLI / Cursor / others: run the engine directly, or copy the usage and cost policy from AGENTS.md into your GEMINI.md or Cursor rules.

Both instruction files describe the same engine and the same cost policy, so behavior is identical no matter which agent drives it.

Headless / autonomous agents

Dev Trace runs cleanly in any headless context (cloud agents, CI, cron) with no code changes:

• Keys come from the environment—set HUNTER_API_KEY, TABSTACK_API_KEY, etc. however your platform injects secrets. No .env file required.
• Non-interactive—no prompts; JSON to stdout, failed layers captured in meta so a run never crashes the task.
• SixtyFour cannot spend without acceptance—the expensive --deep layer refuses to run unless a human has accepted the cost: DEV_TRACE_ALLOW_DEEP=1 in the environment (standing acceptance set by the operator) or an interactive confirmation. An autonomous agent that passes --deep without that opt-in gets a skipped layer logged in meta.errors and zero SixtyFour spend.
• DEV_TRACE_MAX_DEPTH=N caps the default depth (the --max-depth flag still overrides per call), e.g. DEV_TRACE_MAX_DEPTH=2 keeps an agent on the free layers.
• Runtime: needs Node 18+ available in your container/runtime.

Usage

node scripts/dev-trace.mjs <email> [options]

Flag	Effect
--pretty	Human-readable JSON
--json <path>	Write raw JSON to a file
--max-depth N	Stop after depth N (0–4). --max-depth 2 keeps it entirely free
--deep	Add SixtyFour (depth 5). Expensive, rate-limited, opt-in
--tier low|medium|high	SixtyFour research depth (default low)
--no-bio	Skip the Tabstack /research bio
--bio-mode fast|balanced|deep|max	Bio research depth (default fast)

Output

{
  "meta": {
    "email": "someone@example.com",
    "depths_run": [0, 1, 2, 3],
    "match_status": "verified | probable | unverified_candidates",
    "generated_at": "<ISO-8601>"
  },
  "profile": {
    "<field>": { "value": "...", "confidence": 0.0, "sources": ["..."], "conflict": null }
  }
}

• match_status: verified (confirmed + deliverable), probable (found, unconfirmed), unverified_candidates (below threshold).
• Failed or rate-limited layers are recorded in meta.errors / meta.rate_limited.

Principles

• Public data only. No invented profiles. Below the confidence threshold, candidates are returned, not asserted.
• Deterministic. Fixed layer order and flags; output varies only with live source data and which keys are present. meta.depths_run always explains what ran.
• Transparent cost. Free by default; paid layers are opt-in and labeled.

Contributing

Issues and PRs welcome. Keep the engine dependency-free (Node built-ins only) and every new layer key-gated and cost-labeled.

---

MIT © Built for Devs. Hunter, Tabstack, SixtyFour, and GitHub are independent services—bring your own keys and accounts.