Audit Bridge is a Codex skill and scanner that reviews software repositories for privacy, security, and compliance risks, then publishes structured audit findings to AuditBridge.
This project is a submission for the MuShanghai hackathon by the PartyNS team.
- Personal data collected in code paths, schemas, forms, and models
- Data sent to third-party APIs or analytics services
- Logs that may contain emails, tokens, cookies, auth headers, or identifiers
- Auth flows without visible rate limiting
- Cookies,
localStorage, orsessionStorageusage without an obvious consent layer - AI model calls that send user content externally
- Hardcoded secrets, insecure crypto, and missing audit-log coverage
- Open-source dependency license risk
- Missing retention, deletion, export, or data portability paths
- Scans a repository and extracts evidence-backed findings.
- Builds an AuditBridge-compatible JSON payload.
- Posts the payload to
https://auditbridge.partyns.com/api/audits. - Returns a shareable report URL from AuditBridge.
Install this folder as a Codex skill, then invoke it from a repository you want to audit:
Use $auditbridge-compliance-audit to audit this repo. Review the generated findings for obvious false positives, post the final audit to AuditBridge, and give me the final report link.
The bundled scanner can also be run directly:
python3 scripts/auditbridge_scan.py --repo /path/to/repoFor a review-first workflow:
python3 scripts/auditbridge_scan.py \
--repo /path/to/repo \
--no-post \
--payload-out /tmp/auditbridge-payload.json
python3 scripts/auditbridge_scan.py \
--payload-in /tmp/auditbridge-payload.jsonSet AUDITBRIDGE_API_BASE or pass --api-base to use a different AuditBridge endpoint.
SKILL.md: Codex skill instructionsscripts/auditbridge_scan.py: repository scanner and AuditBridge posterreferences/auditbridge-api.md: API payload referencereferences/finding-rubric.md: severity and evidence rubricagents/openai.yaml: Codex UI metadata
Audit Bridge produces heuristic findings to support review. It is not a final legal compliance determination.

