[LIghtJUNction]

Motivation

• The agent state feature permitted unauthenticated clients to create unlimited fresh Ed25519 identities and persist per-identity data, which could exhaust the authoritative SQLite database and WAL because agent_state was not included in global storage accounting.
• The goal is to enforce a global, atomic capacity for persistent agent state so rotating attacker-generated identities cannot bypass site-wide quotas while preserving existing per-slot and per-identity limits.

Description

• Acquire an SQLite write lock (BEGIN IMMEDIATE) before measuring and writing agent_state so concurrent writers cannot race to oversubscribe the limit and use transactional commit/rollback around the insert. (see ExchangeService.state_write in src/msgd/exchange.py).
• Enforce a global capacity check that accounts for per-record overhead and the bytes of owner_id, name, and value, and reject writes with HTTP 507 when the effective capacity would be exceeded; preserve the existing MAX_STATE_SLOT_BYTES and MAX_STATE_TOTAL_BYTES checks. (see src/msgd/exchange.py).
• Expose the effective global agent-state capacity in the rendered schema (agent_state_capacity_bytes) and document the behavior in README.md and src/msgd/render.py.
• Add a regression test (tests/test_msgd.py::LegacyMigrationCase.test_agent_state_has_global_storage_capacity) that proves fresh-identity exhaustion is bounded and that replacing an existing value frees capacity as expected, and update imports to surface StoreError for assertions.

Testing

• Ran static checks with uv run ruff check src/msgd/exchange.py src/msgd/render.py tests/test_msgd.py and git diff --check, which passed without issues.
• Ran focused unit tests with uv run python -m unittest tests.test_msgd.LegacyMigrationCase.test_agent_state_has_global_storage_capacity tests.test_msgd.ExchangeProtocolCase.test_state_watch_ack_task_thread_outbox_and_since and both tests passed (the patch was iterated until the failing interaction was resolved).
• Ran the full test suite with uv run python -m unittest discover -s tests, which completed successfully: 146 tests ran, 1 skipped, all passing.

---

Codex Task

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
🔒 Security Review	✅ Completed 2026-09-28T11:49:34.648819Z	5666772	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[LIghtJUNction]

关闭：现行 src/msg 中不存在 WebSub、持久 agent state、engagement ranking 或 mention 别名扇出，因此没有对应的无界路径可修。此关闭仅表示补丁不适用于现行代码，不代表生产已验收。