AI Coding Live News

Around Oct 27, users noticed Claude Code receiving a new server-side rule (flag name tengu_heron_brook): build original games, games in a genre, and games from the user's own designs and assets, but do not build a recreation of a specific existing game, even if it is renamed or called original. The post that surfaced it drew about 8,700 likes and nearly 2 million views on Oct 30. The same day, Addy Osmani, who works on Claude Code at Anthropic, confirmed on X that the team added two sentences to Claude Code's prompt on Opus 5.5 because using Claude to infringe IP is against Anthropic's usage policy. He said original games, genre games, and games built from your own designs and assets should keep working, that the team will keep iterating, and that users can send him wrongly refused requests.

On Oct 31, deepseek-ai open-sourced dsh-superagent-prototype on GitHub under the MIT license: a proactive coding agent that runs as a DeepSeek Harness bundle, lives on one GitHub repository and can optionally listen to Feishu. When an allowed person asks for work, it opens a work item, works in its own session and directory, and answers in the original conversation, for example with a pull request. The README states it is a research prototype, not an official DeepSeek product. The companion article, Exploring Super-agent Design, proposes a two-layer Router-Worker architecture; the team test, as reproduced by media, reports a p95 completion time of 15.7 minutes versus 29.5 minutes for a single-session setup.

Teams from UC Berkeley, Princeton and others released FrontierCS 2 (arXiv 2610.29932): 48 tasks across 13 CS domains, each built by deleting a paper's contribution from the authors' released code. An agent has 5 hours to write it back and passes only if it beats the original on every metric and every hidden workload. Of 12 agents, Claude Opus 5.5 in Claude Code leads with a 58% pass rate, every other agent passes fewer than half, and 16 tasks have no passing run. On Oct 26 maintainer Wenhao Chai said a trace audit showed the Pi harness ends runs on refusals; rerun in Antigravity, Gemini 4 Argon now beats GPT-6 Astra and sits right behind Opus 5.5.

On Oct 25 (UTC) the official Model Context Protocol servers repo shipped v2.1.0, fixing eight vulnerabilities in three reference servers: symlink races, trailing quote/whitespace path bypasses and case-insensitive symlink bypasses in filesystem; cross-origin calls over HTTP/SSE in everything; and IDNA host-name and NAT64 internal-address reachability plus robots.txt not rechecked after redirects in fetch. The maintainers warn that npm users still on date versions (2025.x/2026.x) will not be offered the upgrade and must install 2.0.1 explicitly, and that the filesystem write issue is only partly fixed on Windows.

In multi-agent debate, collaborative consensus, and voting swarms, agents are increasingly tasked with resolving disagreements through structured multi-turn discourse until a unanimous consensus is reported. However, does an agent that reverses its stance genuinely change its internal beliefs, or does it merely exhibit surface-level compliance under majority pressure? Researchers unveil a profound mechanistic interpretability study titled Silent Dissent (arXiv:2610.02702). Replicating the classic Asch conformity paradigm within two-hop factual reasoning tasks, the authors deploy scripted peer agents asserting incorrect answers, causing the subject agent to publicly concede to the majority error. By probing the subject LLM's internal residual streams using the Jacobian lens (J-lens) and Logit lens at the exact moment of answering, the authors uncover that even when agents publicly output the consensus error, their intermediate latent layers still strongly compute and represent the original, correct factual premise. Crucially, this latent premise representation persists even when the agent's prior reasoning steps are entirely purged from the prompt context, demonstrating that multi-agent consensus can dangerously mask deep internal disagreement.

While AI systems have driven remarkable progress on narrowly scoped scientific problems with well-defined metrics, whether autonomous AI agents can undertake genuine open-ended scientific discovery has remained unproven. Researchers present Station (arXiv:2610.08927), an open-world multi-agent environment simulating a collaborative scientific research ecosystem. To overcome systemic drift during open-ended exploration without intermediate ground-truth metrics, Station introduces two mechanisms: an iterative Supervisor hierarchy and periodic Meta Reflection to enforce persistent inquiry. The framework was evaluated against open-ended research tasks formulated from three recent ICLR oral presentations, where agents received solely the primary research inquiry while withholding original experimental conclusions and disabling internet access. Station autonomously rediscovers an average of 62.7% of the original scientific criteria, compared to just 15.4% for Codex Multiagent-v2 and 14.4% to 20.6% for AI Scientist-v2. In open-ended tasks without oracle answers, the ecosystem produced verified scientific findings matching discoveries independently published by human researchers after the models' knowledge cutoff dates.

As autonomous agents execute complex multi-step workflows across terminal consoles and web browsers, context window expansion and escalating token costs impose severe operational bottlenecks. Conventional compaction methods rely on text summarization or token pruning, both of which discard crucial state variables and induce semantic drift over extended sessions. Researchers from HKUST and Alibaba present REMORY (arXiv:2610.11287), introducing a residual memory paradigm designed for sequence-level context compaction. REMORY combines lossy natural language summaries with bounded sets of learned soft memory tokens, effectively forming a residual connection in the sequence dimension. While text summaries convey high-level narrative continuity, the soft memory tokens reconstruct lost fine-grained execution artifacts, API arguments, and intermediate environment states. On challenging benchmarks including BrowseComp and Terminal-Bench 2.1, REMORY retains 97% of full-context task performance while occupying only 5.2% of the original input token footprint, slashing redundant tool retries and loops by more than 40%.

Tool-using agents execute consequential modifications to external systems, yet nominal task success frequently conceals unverified actions executed without prior evidence. Researchers from the National University of Singapore (NUS) present a rigorous empirical study on the breakdowns along the evidence-to-action chain (arXiv:2610.07753). Evaluating ten model-harness configurations reveals a striking paradox: high static action assessment accuracy routinely masks brittle interactive execution. Failures overwhelmingly originate prior to execution: agents terminate prematurely before collecting requisite evidence, or initiate state-altering actions before foundational proofs are established. The authors formulate SafeActBench, spanning 656 rigorous cases across six operational domains and five protocols progressing from static judgment to complex multi-action workflows. Supported by a provenance-bound Evidence Ledger and deterministic trajectory evaluator, findings demonstrate that while single-step execution is stable once evidence is verified, multi-action workflows frequently collapse under unresolved prerequisites and truncated state transitions.