github-stars-monitor

GitHub Stars & Forks Monitor

Track who stars and forks [YOUR_PRODUCT] repos. Research their profile. Alert when someone from a notable company or with a large following engages.

Why

A GitHub star from a VP Eng at Stripe is worth more than 100 random stars. Forks are even stronger — they indicate intent to evaluate. Cross-referencing GitHub profiles with company data surfaces enterprise prospects hiding in OSS engagement.

Setup

Requires two env vars in ~/.openclaw/gateway.env (or your agent's secrets store):

GITHUB_TOKEN=<fine-grained token with read access to Starring + Contents for the org>
MONITORED_REPOS=owner/repo-1,owner/repo-2,owner/repo-3

Repos to Monitor

Set via the MONITORED_REPOS env var (comma-separated owner/repo entries). Examples:

MONITORED_REPOS=acme-org/acme-cli,acme-org/acme-mcp,acme-org/acme-skills

Workflow

Step 1: Collect New Stars & Forks

Run the collector script. It handles pagination (per_page=100), deduplication against state, and profile enrichment automatically:

python3 /home/openclaw/.openclaw/workspace/skills/github-stars-monitor/collect.py

CRITICAL: Use the exact absolute path above. Do NOT use cd /path && python3 script.py (triggers exec preflight block). Do NOT append 2>&1. Direct form only.

Output is JSON:

{
  "new_stargazers": [{"repo": "...", "username": "...", "starred_at": "...", "profile": {"name": "...", "company": "...", ...}}],
  "new_forkers": [{"repo": "...", "username": "...", "forked_at": "...", "profile": {...}}],
  "repo_stats": {"owner/repo-1": {"total_stars": N, "new_stars": N, ...}},
  "total_new": N
}

State is saved to memory/github-stars.json automatically — processed users won't appear again on next run.

If total_new == 0 → reply HEARTBEAT_OK, done.

Step 2: Qualify

For each new stargazer/forker, score as HOT / WARM / SKIP using profile data from the collector output:

HOT (alert immediately):

company field mentions a known tech company (>100 employees)
followers > 500 (influential developer)
bio contains leadership keywords: VP, CTO, Director, Head, Lead, Principal, Staff
Profile shows they match your ICP profile (active repos, relevant tech stack, contribution patterns)

WARM (include in summary):

Business company in company field but small/unknown
Moderate activity (20+ public repos)
Forked (stronger signal than star)

SKIP:

Empty profile, no company, <5 repos, bot-like

Step 3: Research HOT Candidates

For HOT candidates, run Exa to find more about their company:

exa-search__web_search_advanced_exa \
  query="COMPANY_NAME company" \
  category=company \
  numResults=3 \
  type=auto

If stargazer has no company but has a personal site in blog, check it via Exa for affiliation.

Step 4: Alert

For HOT candidates:

GitHub: [username] starred [repo]
[Real name] — [bio/title]
[Company] — [one-line from Exa research]
[followers] followers | [public_repos] repos
Profile: github.com/USERNAME

Summary for WARM:

Also starred: [N] others ([username1] @ [company], ...)

Repo stats summary:

Stars: [repo1] N (+new), [repo2] N (+new)

Notes

GitHub API rate limit: 5000 req/hour with token. Collector uses ~10-50 requests per run depending on star count.
Fork is a stronger signal than star — forkers appear first in output.
State file keeps up to 2000 processed users per repo. Old entries trimmed automatically.
First run will report ALL existing stargazers as "new". Review and send HEARTBEAT_OK if no HOT candidates.

Related Skills

signup-monitor — if a stargazer later signs up, connect the dots
exa-company-research — deep company research for HOT leads
exa-people-research — deep people research
pipeline-manager — add HOT stargazers as prospects