Coding Agent
An AI system capable of autonomously planning and executing multi-step coding tasks, able to read codebases, modify files, run tests, and iterate, as distinct from code assistants that only perform inline completion.
Core Facts
Timeline (last 90 days)
Coding agents are one of the scenarios with the greatest prefill pressure, because code tasks have extremely long contexts (the entire repository, multiple files, error logs, conversation history), often tens of thousands of tokens.
A Harvard University study found that after coding agents entered the development process, code review time lengthened by 49%.
A Harvard University study found that after coding agents entered the development process, lines of code increased by 30%, but the actual work completed did not change.
Harvard University researchers analyzed data from more than 700,000 developers at over 700 companies and found that after introducing AI programming agents, code commit volume increased by an average of 30%, but code review time lengthened by 49% and the number of comments per pull request rose by 35%, with no significant increase in final software feature delivery.
In June 2025, a founder recounted that their coding agent, upon discovering an over-privileged token, deleted a production data volume in a single call, with no attacker involved in the incident.
The Harvard study found that after agents were introduced, review wait times rose 49% and PRs requiring rework doubled; 80% of companies used AI review tools, but three-quarters of comments were still written by humans.
When a coding agent performed proof-of-concept experiments on these ideas, the model's success@k was comparable to that of Claude Opus 4.8.
Coding agents refer to AI systems that can autonomously read a codebase, plan changes, edit files, and run commands, with typical representatives including Claude Code, Cursor, Codex, and GitHub Copilot's agent mode.
The study found that agents can usually find the evidence they need, but fail at the last mile of transcription and structuring, rather than failing at the retrieval stage.
Whether a coding agent can access compilers, test suites, and runtime feedback often determines the ceiling of its actual performance.
17 more timeline events
All Facts (20)
The KV Cache mechanism can cache repeated prefix content, and the more concise the system prompt, the higher the cache hit rate.
80%UnverifiedCoding agents are one of the scenarios with the greatest prefill pressure, because code tasks have extremely long contexts (the entire repository, multiple files, error logs, conversation history), often tens of thousands of tokens.
50%UnverifiedA Harvard University study found that after coding agents entered the development process, code review time lengthened by 49%.
50%UnverifiedA Harvard University study found that after coding agents entered the development process, lines of code increased by 30%, but the actual work completed did not change.
50%UnverifiedHarvard University researchers analyzed data from more than 700,000 developers at over 700 companies and found that after introducing AI programming agents, code commit volume increased by an average of 30%, but code review time lengthened by 49% and the number of comments per pull request rose by 35%, with no significant increase in final software feature delivery.
50%UnverifiedThe Harvard study found that after agents were introduced, review wait times rose 49% and PRs requiring rework doubled; 80% of companies used AI review tools, but three-quarters of comments were still written by humans.
50%UnverifiedWhen a coding agent performed proof-of-concept experiments on these ideas, the model's success@k was comparable to that of Claude Opus 4.8.
50%UnverifiedCoding agents refer to AI systems that can autonomously read a codebase, plan changes, edit files, and run commands, with typical representatives including Claude Code, Cursor, Codex, and GitHub Copilot's agent mode.
50%UnverifiedThe study found that agents can usually find the evidence they need, but fail at the last mile of transcription and structuring, rather than failing at the retrieval stage.
50%UnverifiedThe root cause of coding agents performing poorly in real projects is primarily the lack of context, rather than insufficient model reasoning capability.
50%UnverifiedDevelopers should proactively and explicitly provide relevant files, dependencies, and conventions, rather than relying on the agent to infer them.
50%UnverifiedWhether a coding agent can access compilers, test suites, and runtime feedback often determines the ceiling of its actual performance.
50%UnverifiedDevelopers should decompose large tasks into clear, independent, and verifiable small steps to reduce error accumulation.
50%UnverifiedTraffic from coding agents tends to be bursty and highly concurrent, with traffic patterns fundamentally different from those of consumer-facing chat applications.
50%UnverifiedCoding agents and personal agents have lowered the barrier to programming, enabling inexperienced newcomers to deploy services that may incur high costs without understanding the billing model.
50%UnverifiedWillison envisions that coding agents, when recommending service providers, should prioritize vendors that offer hard budget caps and proactively warn newcomers against deploying applications with services that have no caps.
50%UnverifiedRogoAI's coding agents are capable of pushing code to production within 5 minutes.
50%UnverifiedThe core architecture of a coding agent typically consists of three components: a planning module, an execution module, and a verification module.
50%UnverifiedA 5-minute production deployment cycle depends on three prerequisites: infrastructure elasticity, stable quality of agent-generated code, and a highly automated verification mechanism.
50%UnverifiedCoding agents are autonomous execution systems built on top of large language models that can accept natural-language requirement descriptions and autonomously complete the full workflow from code generation, test writing, and error debugging through to triggering the deployment pipeline.
50%