Commit Graph
2 Commits
Author SHA1 Message Date
9ec6164f58 fix(security): harden server and service deletion lifecycle (#1220)
* test: TDD regression tests for GHSA-jx78-55p5-rwv5 stream quota enforcement

* Apply remaining changes

* fix: update action SHA allowlist and test assertions to match dependabot bump

* fix: close GHSA-jx78-55p5-rwv5 incomplete fix of GHSA-qjpp-gffx-2wm9

Finding 1 (Moderate): nil-guard reporterServer in delayCheck and notifyCheck.
ServerShared has its own lock independent of serviceResponseDataStoreLock, so
m := ServerShared.GetList() taken inside the worker can return a nil entry for
the reporter if the server was concurrently deleted. Previously this caused an
unrecovered SIGSEGV in the worker goroutine (and in the gRPC layer with no
recovery interceptor), taking down the whole instance.

Finding 2 (Low): nil-guard ss.services[id] in ServiceSentinel.Delete().
A caller-supplied id that is absent from the registry caused
ss.services[id].CronJobID to panic, aborting the Delete loop and leaving every
subsequent valid id as a zombie service (DB row deleted, in-memory entry kept,
cron probe still running).

Regression tests added for both findings following the existing
servicesentinel_lifecycle_test.go patterns.

* Apply remaining changes

* chore: replace commit hashes with version tags in test.yml

* fix(server): serialize authoritative lifecycle changes

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

* fix(service): bind reports to reporter lifecycle

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

* fix(rpc): reject results from stale task streams

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

* fix(agentcompat): allow version-tagged actions

* fix(agentcompat): allow literal checkout refs

* refactor(agentcompat): remove SHA resolver policy

* test(agentcompat): remove resolver SHA fixtures

* test(agentcompat): remove mutable ref fixtures

* test(agentcompat): use tagged actions in secure fixtures

* test(agentcompat): update credential fixtures for tags

* test(agentcompat): update reusable action fixtures

* test(agentcompat): update artifact redaction fixtures

* test(agentcompat): finish artifact fixture tag migration

* test(agentcompat): update workflow validation fixtures

* test(agentcompat): update dependency workflow fixture

* ci(agentcompat): stop pinning cross-repository revisions

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: naiba <hi@nai.ba>
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
2026-08-01 15:45:20 +08:00
naibaandcloudcode e8dabf5bc6 feat(auth): add PAT auth, scoped REST/MCP access, CSRF, and tenant isolation
Introduce Personal Access Tokens (nzp_*) as a stateless auth path alongside
JWT, gated per-endpoint by a scope middleware (nezha:{resource}:{verb}) with
fail-closed empty-scope defaults and a server-id whitelist. Self-management
endpoints (profile, api-tokens, oauth2 bind, refresh-token) explicitly reject
PATs to block privilege-escalation chains. A revoke registry tears down active
long-lived connections (terminal, fm, ws, transfer, mcp) the moment a PAT is
deleted, with a tombstone closing the revoke->register race.

Add an MCP endpoint that proxies tool calls (exec, fs read/write/delete,
transfer) to agents over gRPC, guarded by origin/DNS-rebinding checks, a
per-token rate limiter, audit logging, and a kill switch. Serialize all
sends through the IOStream wrapper to honour grpc-go's concurrency contract.

Add CSRF double-submit protection on unsafe cookie-authenticated methods,
exempting authenticated PAT requests by context identity (not a forgeable
Authorization header). Apply visibility/whitelist filtering consistently
across list, get-by-id, and mutate paths to enforce tenant isolation.

Migrate legacy mcp:* scopes: rewrite read/exec to nezha:* equivalents and
drop dangerous write/delete/wildcard grants.

Co-authored-by: cloudcode <cloudcode@users.noreply.github.com>
2026-05-30 15:56:44 +00:00