Coding Agents

Parent: Agentic.

Coding Agents operate on repositories: plan edits, run tests, open PRs. Under Agentic they are high-signal because git/CI already provide falsifiers. They differ from general Frameworks by assuming a codebase environment and often pair with Code Sandbox for untrusted execution.

Mechanically watch diff quality, test loop integration, permission to write, and how tools expose shell/git. Failure modes: agents that “fix” by deleting tests, unbounded shell, and PRs without reproducible local checks.

Members include Aider, Continue, OpenHands, SWE-agent, and Open Interpreter. Prefer repo-native agents when the job is software change; prefer sandboxes when the agent must execute generated code safely; prefer eval harnesses under Evaluations when you claim SWE progress.

Features such as Coding agents and Self-coding mark the product job: agents that change code with evidence, not autocomplete theater.

Inline members for audit: Aider Continue Open Interpreter OpenHands SWE-agent. Features: Coding agents Self-coding Learning & Adaptation. Related terms referenced: Frameworks, Code Sandbox, Tool Integration.

For this research catalog the acceptance test is boring and strict: after wiring the pack, can a second person replay the same scenario and get the same measurable outcome? If not, the page is still a brochure. Document the scenario, the expected number, and the log line that proves it. That habit is what separates a deepened taxonomy cell from a landscape screenshot.

Also record what you refused: tools not connected, corpora not ingested, metrics not collected. Refusals keep the cell honest when marketing pages claim universality. A taxonomy essay that only lists winners teaches less than one that names the skip conditions next to the install path.

Keep the essay technical enough that an engineer can argue install order — and short enough that a reader finishes it before opening the right-rail member list. Prefer primary docs and repo evidence over landscape blurbs when the right-rail packs disagree.

Operational discipline for this cell: name the install artifact, the health check, and the rollback. Prefer one golden path documented against a member pack in the right rail over five half-wired demos. When packs disagree, believe the primary repository README and a failing test you can re-run — not the landscape one-liner. Measure latency, cost, and error rate on a frozen scenario before declaring the subcategory “covered.” If you cannot point at a log line that proves the scenario, the essay is unfinished even when the word count clears the audit floor.

Cross-link greedily but honestly: every taxonomy wikilink should change a reader’s next click, and every pack slug should resolve to a real pack. Do not invent members to decorate the prose. Prefer fewer true links over many decorative ones. Keep Monday-shaped typing out of the runtime story; observation and ledger remain senior to any agent UI that retypes reality.