Skip to content

Security: wickitlab/Griddy

SECURITY.md

Security

Griddy is a scheduled scraper and static-archive generator. It has no server, no user accounts, and no inbound attack surface of its own; its security posture is mostly about the integrity of the archive and what the scraper touches.

Current posture

What the code touches

  • Network: outbound only, to Enemalta's public outage feeds (live API, planned API, planned HTML page). No inbound listener.
  • Secrets: none. The scraper runs unauthenticated against public feeds and needs no API key or credential.
  • Files: reads and appends to data/*.jsonl and state/current.json, and writes the generated docs/index.html. Nothing outside the repo checkout is touched.
  • Permissions: scrape.yml and ci.yml request only what they need (contents: read, plus contents: write in the scrape workflow to commit new archive rows). No secrets, no external services beyond the three Enemalta sources.

Data-integrity model

Every line in data/*.jsonl carries seq, prev (hash of the previous line), and hash (hash of the line itself), so the whole archive re-verifies from genesis with python -m griddy.verify. Tampering with, reordering, or deleting a line breaks the chain at exactly that point. Git history is a second, independent witness to when each commit landed; the timestamps that matter for a claim are the ones inside the JSONL records themselves.

Personal data from the upstream feeds (InCharge, AffectedAccountNos) is stripped in griddy/normalize.py before anything reaches the archive.

Known weaknesses

  • Records are hash-chained but not yet signed, so integrity is provable but not yet attributable to a specific signing key; ProofLog signing integration is the next roadmap item that closes this.
  • The archive trusts the host clock for ordering between scrape cycles; there is no external trusted timestamp.
  • The scraper depends entirely on Enemalta's public feeds staying reachable and truthful; a feed outage or falsified upstream data would show up as a gap or an incorrect record, not a chain-verification failure.

Reporting a problem

If you find a data-integrity issue (a broken chain, an incorrect archived outage, or a way to make the scraper record something false), please use GitHub's private vulnerability reporting on this repository (Security tab, "Report a vulnerability") rather than a public issue. Include the affected record's sequence number or timestamp and what you expected to see instead.

Disclosure policy

Coordinated disclosure. Once a fix lands, the issue is noted in the changelog.

There aren't any published security advisories