Files
PaperClipAI/doc
Nicky LeachandPaperclip ee81cee76d fix: allow concurrent agent runs on one Anthropic subscription connection (#13445)
## Thinking Path

> - Paperclip manages AI agents and their work.
> - Agent runs use provider connections and subscription credentials.
> - File-backed subscriptions need a lease while a run writes new
credentials.
> - Anthropic subscriptions pass a token and do not write a credential
file.
> - The current lease blocks concurrent Anthropic runs without
protecting shared state.
> - This pull request limits the lease to file-backed subscriptions and
tests both paths.
> - The result lets Anthropic runs share one connection safely while
file-backed credentials remain serialized.

## Linked Issues or Issue Description

**What existing behavior does this improve?**

This change improves concurrent use of one subscription connection.
Anthropic runs no longer wait for a lease that protects no shared file.

**Subsystem affected**

server/ — REST API and orchestration services, with related server tests
and connection documentation.

**Current behavior**

The server takes a connection lease for every subscription invocation.
The lease blocks a second Anthropic run while the first runtime remains
open.

**Proposed behavior**

The server takes the lease only when the subscription writes a provider
authentication file back to the grant. Anthropic runs can proceed at the
same time.

**Reason and benefit**

Anthropic passes its token through an environment variable and writes no
file. Removing this unnecessary wait improves concurrency and preserves
credential safety for OpenAI and xAI.

**Breaking changes**

None. OpenAI and xAI keep the existing lease and retry behavior.

## What Changed

- Limit the AI credential lease to subscriptions that write a provider
authentication file.
- Add coverage for concurrent Anthropic runs and hired-agent runs.
- Keep the existing OpenAI contention assertions as a sensitivity
control.
- Update the AI connection documentation to describe the lease boundary.

## Verification

- Run `server/src/__tests__/ai-connections.test.ts`.
- Run `server/src/__tests__/agent-hire-ai-connections.test.ts`.
- Run
`server/src/__tests__/heartbeat-ai-subscription-contention.test.ts`.
- Run the server type check.
- Review the pull request checks after GitHub starts continuous
integration.

## Risks

The main risk is an incorrect provider classification. The write-back
condition remains unchanged, and OpenAI contention tests keep the
file-backed lease behavior under test. No schema or API contract changes
occur.

## Model Used

OpenAI GPT-5. Exact runtime model ID and context window are not exposed
in this execution. The model used tool calls, shell commands, and GitHub
operations.

## 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>
2026-09-14 20:53:56 -07:00
..
2026-02-16 14:25:00 -06:00