DottaandPaperclip
7ec6e1a237
## Thinking Path > - Paperclip coordinates agents through authorized tools. > - The native runner validates arguments before it dispatches a tool call. > - Schema failures give repair guidance for only five built-in tools. > - Assigned-tool discovery receives a generic rejection, so an agent can mistake invalid arguments for missing access. > - This PR applies the existing safe schema feedback to every authorized tool. > - Agents can correct their calls while authorization, approval, and replay checks remain in force. ## Linked Issues or Issue Description **What happened?** An invalid `paperclip_search_assigned_tools` call returned only `Paperclip rejected this semantic tool call`. The call never reached the assigned-tool gateway. The same feedback gap affects arbitrary assigned app tools. **Expected behavior** Identify argument validation as the cause. Return bounded schema locations and ask the provider to correct its arguments. Do not expose submitted values or grant access to unassigned tools. **Steps to reproduce** Prepare an authorized discovery tool. Submit an invalid query or optional argument through the native runner. Observe the generic error before this fix. Correct the arguments and confirm one dispatch after this fix. **Paperclip version or commit** Reproduced on master at `4970d5ed3d`. The incident did not retain its rejected arguments, so the tests cover representative invalid inputs rather than claim to replay that exact payload. **Deployment mode** Native runner with the Codex app-server driver. Local verification uses macOS ARM64 and controlled provider fixtures. Related: #16868 added bounded monitor validation feedback. #16723 changes runtime MCP catalog loading; this PR changes the Rust runner validation response and does not depend on that work. ## What Changed - Supply generic repair guidance for any authorized tool whose arguments fail JSON Schema validation. Preserve specific built-in hints. - Keep the existing three-location and 512-byte limits. Do not include input values or unknown keys. Preserve generic authorization failures and the existing retry flag. - Cover missing fields, wrong types, null optional fields, bounds, unknown fields, unchanged rejected-call state, corrected calls, and completed-call replay protection. - Exercise rejection, correction, one dispatch, and successful completion through the real Codex driver with a controlled provider subprocess. - Document the behavior and its verification scope in execution semantics and the agent access map. ## Verification - Red: both new runner regressions failed on the original generic rejection. The existing API-validation regression passed. - Green: all seven selected validation tests passed. The new Codex transport regression passed with one rejection and one corrected dispatch. - `cargo fmt --manifest-path packages/paperclip-runner/runner/Cargo.toml --all -- --check`, `node scripts/check-maps.mjs`, and `git diff --check` passed. - Full Rust workspace tests, the repository build, and sequential repository typecheck passed. The existing assigned-tool gateway suite passed all 30 tests. All hosted test and end-to-end jobs passed on this head. The duplicate local `test:run` was interrupted after the hosted suites passed; its local result is incomplete. The hosted build and clean public npm installation check also passed. All current-head CI gates are green, including the final verify and e2e aggregates. - Greptile reviewed commit `3e2743e1e15b3b13a7d803ce4d828510e41fb22a`: 5/5, no actionable findings. - Setup evidence: master has a stale dependency lockfile. Dependencies were resolved with the pinned pnpm 9.15.4; lockfile changes are excluded. The first full typecheck overlapped the build and failed copying a generated shared-package file. The sequential rerun passed with no source change. The initial copy failure remains recorded here; the concurrent build-output race was not separately reproduced. - The provider is simulated. This proves runner validation and transport recovery, not that a live model always retries or that GitHub credentials work. ## Risks The provider now sees schema locations for all authorized tools. The existing sanitizer and bounds still apply. Unknown input keys and values remain excluded. No schema, permission, approval, automatic tool replay, database, or connection-auth changes are made. Unauthorized tools still get a generic rejection. ## Model Used OpenAI Codex (GPT-6), with reasoning, source inspection, and tool execution. The exact runtime model ID and context-window size are not exposed in this session. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge Co-authored-by: Paperclip <noreply@paperclip.ing>
8.5 KiB
Tool gateways and access profiles — agent access
Product map: Tool gateways and access profiles.
Review date: 2026-10-12. Source commit: 58e16dbddfc9c8d0cf60867f9d9847f0feda055b.
Native REST transport reconciled in this change; see the
shared transport contract.
This is a source-reviewed baseline of representative operations and known gaps. Each capability needs operation-level review before its entire scope can be qualified. No live agent execution is claimed. Follow the map contract for format, permission checks, evidence and maintenance.
gateway
Review: baseline.
API: present — representative operation IDs:
POST /api/companies/{companyId}/tools/gateways.
Agent access: restricted — Board tools:admin REST; CLI connections/configuration where
supported.
Policy: Named gateway/profile administration and credentials are board tools:admin; agent sessions/tool use use separate scoped authority.
Evidence: source-only — implementation
at the reviewed commit above.
Verification: not-run — assert the documented agent denial with an owned test fixture;
verify any listed agent request/result path separately. A delegated operation needs its
own policy and access review.
Gaps: No agent named-gateway management path; assigned tool use exists separately.
profiles
Review: baseline.
API: present — representative operation IDs:
POST /api/companies/{companyId}/tools/profiles.PATCH /api/tool-profiles/{profileId}.
Agent access: restricted — Board management REST; agent receives effective tool
assignment.
Policy: Named gateway/profile administration and credentials are board tools:admin; agent sessions/tool use use separate scoped authority.
Evidence: source-only — implementation
at the reviewed commit above.
Verification: not-run — assert the documented agent denial with an owned test fixture;
verify any listed agent request/result path separately. A delegated operation needs its
own policy and access review.
Gaps: Agent access to tools does not permit profile editing.
Backend evidence: profile entry validation covers company-scoped selector references and bounded reads for a large immutable runtime assignment. Batched validation retains the checks before profile and entry writes. It grants no profile-management authority to agents.
client-connect
Review: baseline.
API: present — representative operation IDs:
GET /api/tool-gateway/gateways/{gatewayId}/mcp.POST /api/tool-gateway/sessions.
Agent access: partial — Agent REST / native call_api for scoped session creation
and revocation; existing MCP client protocol and assigned native app tools.
Policy: Named gateway/profile administration and credentials are board tools:admin; agent sessions/tool use use separate scoped authority.
Evidence: source-only — implementation
at the reviewed commit above.
Verification: not-run — exercise a listed permitted operation with actual agent
credentials/run context; inspect its result and applicable persistence/attribution.
Assert the relevant denied actor, owner, company or mode from Policy. Review the
remaining operations before qualifying the whole capability.
Gaps: Scoped sessions remain agent/run-bound. MCP requires its existing protocol client; named gateways and gateway-token management retain board permissions.
Discovery collapses identical descriptors with the same gateway name across providers, including native assigned-tool startup and fresh searches. Conflicting definitions remain an error. Distinct names and the normal invocation grants, approvals and revocation checks remain independent. Regression coverage lives in gateway tests and assigned-tool tests; these fixtures do not certify a live external-client journey.
Native assigned tools return bounded input-schema repair guidance after argument validation fails. Rejection creates no execution receipt and discloses no submitted values or unknown keys. Corrected calls keep the same authorization checks. Runner regressions cover discovery repair and unauthorized calls; the Codex transport fixture covers rejection, corrected dispatch, and completion with a simulated provider. This does not establish live model or GitHub access.
Remote invocation regressions preserve access to authorized sibling tools after a fetch failure, timeout, non-authentication HTTP error, or JSON-RPC rejection. HTTP 401/403 and insufficient-scope errors keep their connection gates. The gateway records each failed invocation and never automatically replays it. This covers the gateway service with synthetic providers and a local HTTP server, not a live external provider or all client surfaces.
tokens
Review: baseline.
API: present — representative operation IDs:
POST /api/tool-gateway/gateways/{gatewayId}/tokens.POST /api/tool-gateway/gateway-tokens/{tokenId}/revoke.
Agent access: restricted — Board tools:admin token management.
Policy: Named gateway/profile administration and credentials are board tools:admin; agent sessions/tool use use separate scoped authority.
Evidence: source-only — implementation
at the reviewed commit above.
Verification: not-run — assert the documented agent denial with an owned test fixture;
verify any listed agent request/result path separately. A delegated operation needs its
own policy and access review.
Gaps: Gateway-token management requires board tools:admin at the HTTP route.
A grant to use an app does not create this authority; native call_api preserves
the denial.
activity
Review: baseline.
API: present — representative operation IDs:
GET /api/tool-gateway/audit.
Agent access: restricted — Board tools:view_audit; agent receives its own tool
receipts/errors.
Policy: Named gateway/profile administration and credentials are board tools:admin; agent sessions/tool use use separate scoped authority.
Evidence: source-only — implementation
at the reviewed commit above.
Verification: not-run — assert the documented agent denial with an owned test fixture;
verify any listed agent request/result path separately. A delegated operation needs its
own policy and access review.
Gaps: Audit management API exists but broad audit visibility is not an agent use-path guarantee.
Execution-host stdio assignments
Runner registers verified, session-local stdio servers from the execution host with native Codex and through the shared ACPX runtime. ACPX clients must support stdio MCP assignments. The transport uses each implementation's own schemas and responses. Each binding requires a private Unix socket to identify the shared resource whose calls are serialized. This is a separate binding alongside the HTTP gateway; it does not create gateway/profile administration authority. Tasks cannot supply these reserved environment inputs. Read-only sessions omit them. ACPX adapters use their applicable ACPX permission policy. Host-assigned native Codex tools run without a new per-call prompt, matching existing execution-host shell utilities. Explicit deny/prompt rules remain effective; requested approvals use PRP with active thread/turn checks and one-time decisions. Preconfigured Claude rejects interactive approval requests. See the Runner contract.
Local fixture coverage: binding validation, relay processes and cancellation, and approval projection. Provider-client image and approval smoke tests used native Codex and Claude SDK; that coverage is not an exhaustive adapter compatibility test. Vendor-client evidence belongs in the PR. These checks do not certify live external credentials, deployment isolation or gateway revocation.