dpdp-cli

Read-only DPDP compliance scanner for the DPDPOS Compliance Management Platform.

It collects evidence from code/config/docs, packages structured findings, and ships them to the backend. It never mutates customer source files.

Install from npm

Requires Node.js 20+.

npm install -g dpdp-cli

Or run without a global install:

npx -p dpdp-cli dpdp <command>

The published binary name is dpdp.

Publishing this package (maintainers)

1. Set "private": false in package.json when ready.
2. npm run build then npm test.
3. npm login → npm publish --access public.

Point end users at your deployed API URL (https://…), not localhost.
Full stack deploy (API, worker, Vercel, Neon, Upstash, R2, Entra, npm):
see dpdpos_backend/docs/14_deployment.md.

Directory identity (Windows AD / Entra / Microsoft 365)

Enterprise SSO does not change how the CLI authenticates for scans.

1. Operators sign into the frontend with Microsoft Entra, Windows AD (LDAP), or local password.
2. They mint a dpdp_… assessment CLI token in the Assessments → CLI tab (requires assessment:cli_token).
3. The CLI continues to use that opaque token for scan / submit / status.

This keeps machine evidence tokens separate from human SSO sessions (better audit, shorter blast radius if a laptop token leaks).

Optional: DPDP_USER_TOKEN (user JWT after SSO) may still be used only for the optional report command.

Where to get the CLI token

Use the product frontend (not PowerShell, not /demo):

1. Open the DPDPOS console and sign in (password, Microsoft, or Windows AD)
2. Go to Assessments → open (or create) an assessment
3. Complete Documents + Questionnaire tabs
4. Open the CLI tab → Generate CLI token
5. Copy the one-time dpdp_… token and the command block (includes npm install -g dpdp-cli and the production --api URL)

Create a new assessment version in the Overview tab before a historical rescan. Then run scan + submit again on that version.

Evaluation and reports live in the frontend Results tab (Evaluate controls). The CLI report command is optional and needs a separate user JWT (DPDP_USER_TOKEN).

Commands

npm install -g dpdp-cli

dpdp init
dpdp login --token <dpdp_...> --api https://your-api.example.com
dpdp configure --assessment <uuid>
dpdp scan .
dpdp evidence
dpdp submit
dpdp status

Local development (from this repo)

npm install
npx tsx src/index.ts login --token <dpdp_...> --api http://127.0.0.1:3000
npx tsx src/index.ts configure --assessment <uuid>
npx tsx src/index.ts scan ./fixtures/sample-app
npx tsx src/index.ts evidence
npx tsx src/index.ts submit
npx tsx src/index.ts status

# VAPT capability (passive, authorized scope only; backend not connected yet)
npx tsx src/index.ts vapt scope --target https://app.example.com --target-type URL --authorized-by you@corp --purpose "release assessment"
npx tsx src/index.ts vapt scan
npx tsx src/index.ts vapt findings

Notes

• scan saves findings locally first, then creates a remote scan job.
• Token values from the platform always start with dpdp_.
• rescan = scan + submit on the current assessment version (version bump is done in the frontend).
• vapt is a passive, non-destructive capability: it refuses to run without an explicitly authorized scope, never contacts excluded targets/ports, and stores findings locally under ~/.dpdp/. Backend VAPT APIs do not exist yet, so vapt submit/vapt status report local state and show the exact payload that will be submitted once connected.

Local storage

Local state lives under ~/.dpdp/ and is migrated automatically from the legacy single-file ~/.dpdp/config.json on first use (non-destructive):

~/.dpdp/
  schema.json                    storage schema version
  config/config.json             apiBaseUrl, assessmentId
  config/vapt/<assessmentId>.json  per-assessment VAPT scope (Phase 4)
  credentials/credentials.json   CLI token (mode 0600 on POSIX)
  state/current-scan.json        pointer to the current scan (DPDP only)
  state/scans/<scanId>.json      per-scan state (job id, status, timestamps; VAPT scans carry capability + extra)
  evidence/<scanId>.json         per-scan evidence (DPDP findings, or VAPT findings at schemaVersion 2)

Evidence is stored per scan, survives failed submissions, and can be retrieved offline; the bearer token never appears outside credentials/credentials.json.

Development

npm install
npm run build   # typecheck + compile to dist/
npm test        # node:test via tsx (no extra test deps)

The CLI is organized as an evidence-collection platform: the scanner engine (src/core/scanner/) orchestrates collectors → analyzers → evidence normalization/dedup. The current regex scanner is one analyzer (src/analyzers/source/regex/); future analyzers register into the engine via src/core/profiles/ without touching the engine or CLI entry point.

Design documents

The data and execution contract for the future VAPT capability is defined in docs/vapt/ (design only — nothing implemented).

Related repos

| Repo | Purpose | |---|---| | dpdpos_backend | API + worker — see docs/14_deployment.md | | dpdpos | Next.js console (Vercel) |