Context compaction in coding agents can launder forged user instructions
Flattening harnesses accept forged approvals 87-100% of the time, while four role-typed harnesses reject the same attack.
58
Product
OpenHands
- Mentions · 7d0
- Mentions · 30d2
- All mentions5
- First seenSep 18, 2026
- Last mentioned19d ago
Timeline
5 articles · newest firstHarnessAdapt-RL trains coding-agent policy and harness together
On a 9B model, pass@1 rises 9.8, 12.6, and 6.4 points over GRPO on three SWE-bench sets.
52
Opera: A Verbal Critic Framework for Long-horizon Coding Agents
Opera, a persistent verbal critic, lifts coding-agent resolve rates by up to 15 points on SWE-Bench-style benchmarks.
52
SkillSeek: Revisiting Agent Skill Retrieval at Marketplace Scale
SkillSeek shows standard retrieval matches LLM-mediated agent skill selection at about half the cost.
44
Best Open-Source Agent Harnesses for Local LLMs in 2026
A 2026 guide ranks 11 open-source agent harnesses, including OpenCode, Goose, Cline, and OpenHands, by local-LLM support, licensing, and safety controls.
28
Appears with
- opencodeproduct · 2
- Aiderproduct · 2
- Clineproduct · 2
- piproduct · 2
- HarnessAdapt-RLproduct · 1
- SkillSeekproduct · 1
- Continueproduct · 1
- Qwen Codeproduct · 1
- SkillsBenchproduct · 1
- Ollamaproduct · 1
- Gooseproduct · 1
- Anthropicorganization · 1
- SWE-Bench Proproduct · 1
- Terminal-Bench 2.1product · 1
- DeepSWEproduct · 1
- Gemini CLIproduct · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.