chore(maintenance): refresh dependencies and resolve security findings - #2378
Merged
Conversation
Deploying with
|
| Status | Name | Latest Commit | Preview URL | Updated (UTC) |
|---|---|---|---|---|
| ✅ Deployment successful! View logs |
bashkit | bf4f394 | Commit Preview URL Branch Preview URL |
Sep 05 2026, 05:50 PM |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What changed
Maintenance now means analyze, fix, validate, and ship through green CI and merge. The skill, command, and spec agree; workflow skills are internal and synchronized globally.
Resolve the confirmed DeepSec findings: preserve complete browser saves, restore supplied JS snapshot callbacks, cap escaped Anthropic output, implement the current Deep Agents backend protocol, fix SQLite setup, include every WASM job in the CI gate, isolate example and publish credentials, and build binary releases from the exact validated commit. WASM tooling derives its exact version from Cargo.lock and replaces stale caches. Release examples use the reviewed lockfile, and CI drops unused write permissions and persisted checkout credentials.
Refresh Rust/npm dependencies, review supply-chain deltas, and document tested security constraints. Monty/get-size2 remain exactly pinned to preserve per-VM limits. The unused jiter Python FFI feature is banned after source review; no new audit exemptions or publisher trust were added.
Why
Maintenance previously stopped at reports and deferral issues. Existing tests also missed adapter, persistence, and release-boundary failures. This pass fixes and verifies those paths end-to-end.
Before / After
Browser traversal failures previously overwrote good saves; three failing regressions now pass, including binary/unreadable entries and deleted roots.
Keyed/unkeyed BashTool snapshot constructors lost callbacks; both regressions now pass. Full JS suite: 573 passed.
Deep Agents integration failed against its current protocol; 831 Python tests pass, with four Linux-only skips on macOS. A real
create_deep_agenttest invokes file/search/execute tools through a deterministic model.Credential/tag/SQLite/CI-gate regressions failed before fixes; 77 script tests now pass, including actual shell execution, moved-tag Git fixtures, and compiler environment checks.
1,859/1,859 Linux/GNU Bash parity cases pass, along with 5,564 Rust workspace tests, 167 rustdoc tests, failpoint/security property suites, 63 WASM tests, and browser/CLI/Python/TypeScript smoke tests.
just pre-pr, all-feature clippy/rustdoc, advisory/license checks, source audits, workflow lint/parity, fuzz compilation, and site/generated-document checks pass. Coreutils regeneration produces no drift.Parallel/SQLite Criterion baselines are saved. The comparison harness now selects and records a supported Bash from
PATH: 96/96 cases match with zero errors against GNU Bash 5.3.15; seven harness tests pass.Closes #2372.
Risk
Medium: broad dependency refresh and an intentional move to Deep Agents >=0.7.13 (Python >=3.11 for that optional integration). Tested exact pins preserve Monty's security contract. Credential isolation protects execution process environments on trusted runners; it is not a sandbox against dependencies persisting across steps. Performance baselines are observational, not controlled speedup claims.
Checklist