key-cape/SCOPE.md

138 lines
9 KiB
Markdown
Raw Normal View History

# SCOPE
Reviewed 2026-09-05 against source revision `b989de4`.
## Purpose and boundary
KeyCape is Go identity **Tooling** for NetKingdom's lightweight IAM deployment.
It runs its own OIDC-style issuer, uses Authelia for upstream browser
authentication, reads identities from LLDAP, and delegates factor checks to
privacyIDEA. It also provides caller authentication commands, directory snapshot
validation, and migration artifact generators.
It issues identity, role, scope and assurance claims. Resource authorization
belongs to access-engine and consuming services; secret custody and OpenBao
policy/token enforcement belong to their platform owners. The implemented subset
supports NetKingdom integrations, but complete profile conformance and drop-in
Keycloak interchangeability are not established.
## Implemented capabilities
| Surface | What the current source provides |
| --- | --- |
| Issuer HTTP API | Discovery, `/authorize` and its callback/registration-return routes, `/token`, `/jwks`, `/userinfo`, local `/logout`, and `/healthz`. |
| Human authentication | Authorization code with S256 PKCE, exact registered redirects, scope allow-lists, Authelia login, privacyIDEA MFA challenges, client-specific assurance requirements, browser session reuse, step-up/fresh-login handling, and configured registration/enrollment handoffs. |
| Service authentication | Static confidential `client_credentials` clients authenticated with form-encoded `client_secret_basic`; configured subject, tenant, roles, scopes and per-client token lifetime. Secrets are resolved from environment references at startup. |
| Tokens and identity | Locally signed RS256 JWTs; configurable access-token resource audience while ID tokens retain the client audience; human/service principal types, tenant, groups, roles, scope and assurance claims. UserInfo resolves canonical directory subjects and filters profile/email/groups by scope. |
| Caller commands | `keycape login` for public-client browser PKCE login and `keycape service-token` for service exchange. HTTPS discovery/JWKS verification and private JSON token-file delivery outside Git; no token output on stdout. |
Make the LLDAP export report its own completeness The exporter discovered groups by walking each user's memberships, so a group nobody belongs to never reached the snapshot, and a failed lookup was skipped by a `continue` under a comment claiming it was recorded in the incompatibility report. The run then emitted `result: "success"`. Add an optional `domain.GroupLister` capability and implement `ListGroups` on the LLDAP adapter as a direct group-subtree search, kept off `UserRepository` because the OIDC layer never enumerates the directory. Record `groupEnumeration` on every result and a `Complete()` predicate over it; abort rather than write a smaller snapshot when the enumeration fails; report a failed per-user lookup on the fallback path; emit `partial` telemetry and name the mode from the CLI. Reading the adapter to write this surfaced a defect the assessment had not listed: `LookupGroups` never populated `Group.Members`, and the exporter built every membership from that field, so against a real directory the `memberships` block was always empty while the fixture-backed tests passed. Memberships on the fallback path now come from the user/group pair actually observed. Sort users, groups and memberships so an unchanged directory exports identically. Closes G05 of the scope/intent assessment. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012WAsfsfQmDu4vcBhiMcmQp Assistant: claude-code Assistant-Model: opus Assistant-Process: 867844@bnt-lap001 Assistant-Session: 3d45905e-0016-4b49-b828-231406881f7b
2026-09-07 08:45:50 +02:00
| Validation and migration | Canonical snapshot checks; deterministic LLDAP user/group/membership export that records whether it enumerated the whole directory; basic Keycloak realm JSON; LDIF generation for OpenLDAP, 389 Directory Server and AD targets. These generate artifacts rather than execute a complete migration. |
| Diagnostics and packaging | Structured authentication/enforcement/migration events, a process health response, Go build/test/vet targets, a container containing the KeyCape binary, and development/CI scaffolding. |
## Material limits
- Client registrations, service secrets and the signing key are loaded at startup.
Login, authorization and handoff state are process-local. There is no shared
session store, general refresh-token flow, token introspection/revocation API,
or automatic signing-key/client-secret rotation service. Logout clears the
local KeyCape session, not every upstream or downstream session/token.
- The optional tenant-engine `tenant_roles` adapter and handler support exist,
but the server entry point does not configure them. That claim is not an
enabled capability of the stock executable.
Make the LLDAP export report its own completeness The exporter discovered groups by walking each user's memberships, so a group nobody belongs to never reached the snapshot, and a failed lookup was skipped by a `continue` under a comment claiming it was recorded in the incompatibility report. The run then emitted `result: "success"`. Add an optional `domain.GroupLister` capability and implement `ListGroups` on the LLDAP adapter as a direct group-subtree search, kept off `UserRepository` because the OIDC layer never enumerates the directory. Record `groupEnumeration` on every result and a `Complete()` predicate over it; abort rather than write a smaller snapshot when the enumeration fails; report a failed per-user lookup on the fallback path; emit `partial` telemetry and name the mode from the CLI. Reading the adapter to write this surfaced a defect the assessment had not listed: `LookupGroups` never populated `Group.Members`, and the exporter built every membership from that field, so against a real directory the `memberships` block was always empty while the fixture-backed tests passed. Memberships on the fallback path now come from the user/group pair actually observed. Sort users, groups and memberships so an unchanged directory exports identically. Closes G05 of the scope/intent assessment. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012WAsfsfQmDu4vcBhiMcmQp Assistant: claude-code Assistant-Model: opus Assistant-Process: 867844@bnt-lap001 Assistant-Session: 3d45905e-0016-4b49-b828-231406881f7b
2026-09-07 08:45:50 +02:00
- The LLDAP export enumerates the group subtree directly, so groups with no
members are present, and every snapshot carries a `groupEnumeration` field
saying whether that enumeration ran or the membership-derived fallback did
(KEY-WP-0018). A failed enumeration aborts the export rather than writing a
smaller snapshot, and read failures are reported rather than skipped. Users,
groups and memberships are sorted, so an unchanged directory exports
identically. Completeness beyond users, groups and memberships — passwords and
MFA credentials in particular — is still not covered.
Make the Keycloak transform preserve or name every policy field keycape-to-keycloak called Transform, which passes no clients, so it wrote a realm with an empty clients array and nothing said the service-identity contract had not been migrated. Where clients were supplied, mapClient hardcoded standardFlowEnabled — silently giving every client_credentials registration the browser flow — and dropped audience, service subject, tenant, roles, lifetime, MFA policy, secret reference and handoff URLs. Realm roles and client scopes were emitted as empty containers. The defect was not the missing mapping but that a dropped field and an inapplicable one looked identical in the output. Add -clients, reading registrations through a new config.Registrations() that converts without resolving secrets, so migration tooling cannot load material it has no business holding. Derive flows from the declared grants. Carry the profile claims as protocol mappers, since Keycloak has no native concept for them, and lifetime, handoff URLs and the secret reference as attributes — the reference, never a value. Derive realm roles and client scopes from what is present. Report what cannot be carried, in UnpreservedReport, kept deliberately separate from ValidationReport: consistency with the snapshot and completeness of the migration are different questions and one list cannot answer both. It names the unmigrated secret, the unenforceable MFA policy, passwords and factor enrolment, and subject continuity. An incomplete transform emits partial telemetry. Closes the semantic-preservation half of G03; proof against a live provider is G04 and stays open. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012WAsfsfQmDu4vcBhiMcmQp Assistant: claude-code Assistant-Model: opus Assistant-Process: 867844@bnt-lap001 Assistant-Session: 3d45905e-0016-4b49-b828-231406881f7b
2026-09-07 13:48:48 +02:00
- The Keycloak CLI takes a `-clients` config and migrates client registrations:
flows follow the declared grants, and audience, tenant, service subject and
roles are carried as protocol mappers with per-client lifetime, handoff URLs
and the secret *reference* as attributes (KEY-WP-0020). What it cannot carry it
names, in a report separate from the consistency check: secret values, MFA
policy enforcement, passwords and factor enrolment, and subject continuity —
Keycloak mints its own `sub`, so the LLDAP canonical ID survives only as an
attribute. The output is a reviewed artifact, not a proven migration; that
proof needs a live provider swap and is not established.
- Snapshot validation is a limited Go rule set, not full machine-readable schema
2026-09-07 00:22:49 +02:00
enforcement. The canonical YAML model and discovery metadata now match the
runtime client-registration surface and are held there by a conformance check
(KEY-WP-0017); the Go model is the runtime authority and the YAML the reviewed
contract. That is narrower than schema enforcement in general.
- The authorization-code grant now binds the redirect URI,
Harden the authorization-code grant and UserInfo verification Closes the local protocol surface of gap G01 from the scope assessment (KEY-WP-0016). The browser grant validated PKCE, client id and scopes but left four bindings unenforced, and UserInfo verified less than the caller CLI does. Authorization-code path: bind the exchange to the redirect URI the code was issued for, refuse clients whose registration does not permit the grant, and authenticate confidential clients with a digest-based constant-time comparison over the same credential sources as the service grant. An empty grantTypes stays an implicit authorization-code client, matching config validation. Code consumption: SessionStore.Consume reads and deletes under one lock. The previous Get/Delete pair spanned JWT signing, and the added test reproduces the race against that version -- 9 of 16 concurrent exchanges succeeded, and a failed exchange left the code replayable. UserInfo: check the JOSE header algorithm before trusting the signature, require the configured issuer, and require an access token rather than accepting an ID token of the right shape. Purpose is decided on the scope claim so the issued token contract, which consumers pin exactly, does not change. SCOPE.md and the assessment record which bindings are now enforced and that the Authelia upstream-trust assumption remains open, so G01 is not fully closed and no profile-conformance claim is made. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NV9oijZukGyGbRQGGKnK4P Assistant: claude-code Assistant-Model: opus Assistant-Process: 713576@bnt-lap001 Assistant-Session: 384c511d-9bce-4cb8-a676-2aef6c0c8df6
2026-09-06 22:43:47 +02:00
enforces grant-type eligibility, authenticates confidential clients and
consumes codes atomically, and UserInfo enforces algorithm, issuer and
Verify upstream Authelia ID tokens Closes the remaining half of gap G01. The adapter decoded upstream ID-token claims without verifying anything, justified in a comment by a server-to-server TLS boundary that nothing enforced. Operator decision: verify the token rather than police the transport. The hop is to be HTTPS as defence in depth, but KeyCape does not monitor, check or gate on that -- a transport check helps only when it is configured correctly, which is the assumption it was meant to remove. Verification holds regardless of how the token arrived, so no HTTPS validation or opt-in flag is added. HandleCallback now verifies the RS256 signature against Authelia's published keys, the issuer Authelia advertises, KeyCape's own client ID in the audience, and a sane validity window, before any claim is trusted. It fails closed: an unreachable or unparseable key set denies the login. The advertised jwks_uri path is rebased onto the server-side token base URL so split-horizon deployments resolve, with config overrides where that inference is wrong, and an unknown key id triggers one refresh so provider rotation needs no restart. The reusable half lives in internal/jose rather than being copied from authclient's verifier, since duplicated verification is how two copies drift and one misses a fix. Migrating authclient onto it is tracked as KEY-WP-0019-T05, kept separate so it does not destabilise a tested path in this change. Thirteen rejection cases plus algorithm and rotation coverage; with the unverified parse restored all fifteen fail, so they test the fix rather than merely passing. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NV9oijZukGyGbRQGGKnK4P Assistant: claude-code Assistant-Model: opus Assistant-Process: 713576@bnt-lap001 Assistant-Session: 384c511d-9bce-4cb8-a676-2aef6c0c8df6
2026-09-07 08:51:42 +02:00
access-token purpose (KEY-WP-0016). Upstream Authelia ID tokens are verified
before any claim is trusted — signature against the provider's published keys,
advertised issuer, own client ID in the audience, validity window — failing
closed and independently of transport (KEY-WP-0019). KeyCape deliberately does
not check or gate on the upstream transport. Enforcing these bindings
is not complete profile conformance. Relying parties must repeat `redirect_uri`
on the token exchange and present the access token, not the ID token, to
`/userinfo`; see [authorization-code bindings](docs/authorization-code-bindings.md).
See the assessment below.
- Tests cover local handlers, adapters, transformations and CLI protocol behavior.
They do not establish complete replacement against a running Keycloak/full-LDAP
stack. The Scenario B/C shell harnesses are incomplete.
- The server listens on HTTP; HTTPS termination is deployment-owned.
`/healthz` reports process status without probing dependencies. Development
Compose needs configuration/key material absent from the checkout. Production
deployment and custody are external; a source implementation or example client
fragment does not prove live registration or consumer cutover.
- Approval-client provisioning, coordinated Qonto rotation, native consumer
handoffs and notification receipt closure remain open in KEY-WP-0009/0013/0014.
## Deliberate exclusions
General-purpose IAM, dynamic client registration, implicit/password grants,
wildcard redirects, arbitrary identity brokering, resource authorization policy,
OpenBao credential custody, and Keycloak operations are outside this repository's
ownership. Registration/enrollment handoffs do not implement user provisioning
or factor enrollment themselves.
## Entry points and verification
Run the issuer with `keycape --config PATH`; `keycape server` and
`keycape migrate` are not implemented subcommands. Caller commands are
`keycape login` and `keycape service-token`.
Separate binaries under `src/cmd/` are `validator`, `lldap-export`,
`keycape-to-keycloak`, and `lldap-to-ldap`. `make -C src build` places binaries in
root `bin/`; the container packages only `keycape`. Tests live under `src/tests/`
and alongside packages in `src/internal/`.
`make test`, `make lint` (Go vet), `make build`, and `make contract-test` are the
root checks. The capability-contract check requires the sibling NetKingdom
validator. These checks passed for the reviewed revision in the preceding
implementation session; the scope assessment is a source/documentation review.
Use this repo for its bounded issuer, caller JWT acquisition, identity adapters
and migration preparation. Consult [INTENT.md](INTENT.md) for the architectural
target, [native authentication](docs/native-authentication.md) for caller
constraints, and the [scope/intent assessment](history/2026-09-05-011726-scope-intent-assessment.md)
for evidence, priorities and remaining gaps.
## Provided capabilities
```capability
type: security
title: Bounded OIDC and service-token issuance
description: Implements static-client browser PKCE and client-credentials authentication with RS256 identity claims through Authelia, LLDAP and privacyIDEA; complete Keycloak interchangeability remains unproven.
keywords: [oidc, pkce, authentication, iam, sso, mfa, identity, service-token]
```
```capability
type: security
title: Verified caller JWT acquisition
description: Provides native browser login and service-token exchange commands with HTTPS discovery, JWKS verification and private token-file delivery.
keywords: [oidc, pkce, jwt, cli, authentication, jwks]
```
```capability
type: security
title: Directory validation and migration artifact generation
description: Checks canonical snapshots and produces LLDAP exports, basic Keycloak realm JSON and target-specific LDIF; does not perform or prove a complete live migration.
keywords: [migration, identity, lldap, keycloak, ldif, validation]
```