Skip to content

Add Harness Engineering for Language Agents (CAR) to Essential Readings - #41

Open
ChaoYue0307 wants to merge 1 commit into
Picrew:mainfrom
ChaoYue0307:add-harness-engineering
Open

Add Harness Engineering for Language Agents (CAR) to Essential Readings#41
ChaoYue0307 wants to merge 1 commit into
Picrew:mainfrom
ChaoYue0307:add-harness-engineering

Conversation

@ChaoYue0307

Copy link
Copy Markdown

Adds "Harness Engineering for Language Agents: The Harness Layer as Control, Agency, and Runtime" (He et al., Preprints 2026) to Essential Readings & Ecosystem Maps.

A position paper that treats the harness layer as a research object: proposes the control–agency–runtime (CAR) decomposition, audits 63 harness-relevant works across academia and public engineering notes, and introduces HarnessCard for structured harness reporting — an academic complement to the practitioner readings in this section.

Followed the maintenance workflow: edited data/projects.yaml (alphabetical, bilingual summaries, tags), ran render_readme.py (README.md + README_zh.md regenerated) and verify_catalog.py (report included; 0 broken URLs — the one remaining stale-github flag is a pre-existing entry).

Note on the link: the canonical page is https://www.preprints.org/manuscript/202603.1756/v2 (DOI 10.20944/preprints202603.1756), but preprints.org returns HTTP 403 to automated checkers (incl. verify_catalog.py), so the entry links the paper's Sciety activity page, which resolves cleanly and carries the DOI. Happy to switch to the preprints URL if you prefer to allowlist it.

Disclosure: I am an author.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant