Topic
Structured skills, prompts and step sequences used to direct AI coding agents through planning and implementation.
sdlc-playbooks runs a gate script before each of its seven coding-agent phases, and a failed precondition aborts the Claude Code command with exit code 2. The check binds only at that entry point, so direct script calls and other editing paths still need their own controls.
Reality
Jarred Sumner's four-month rewrite checked transpiled Rust against a TypeScript suite that does not depend on the implementation language, and reaching a fully passing run cost about $165,000 in tokens.
Reality
A consultant's six-month log across forty shipped Python automations reports build time 38 percent below baseline and client revisions halved, with an explicit warning that React work sits outside the sample.
Reality
SWC's creator fanned a single instruction out to 45 concurrent Labor0 tasks on the ES minifier, then kept review and merge for himself. He says he does not know how long the run took; he was looking at his phone.
Reality
In one developer's fully automated implement-and-review loop, a 1,000-line feature came back at over 4,000 because reviewer findings needed no evidence. Forcing each finding into a failure scenario cut what the implementer had to fix.
Reality
Three postures exist for an agent instruction file. Only the third holds at commit speed. The team running it watched its own CLAUDE.md reach 548KB before a context-measurement pass cut it to 34KB.
Reality
A dev.to guide proposes attaching intent, a risk rating, evidence and rollback steps to every AI-assisted pull request. The template holds up as far as it goes, but it never asks the agent which files it actually read.
Reality
An embedded loop stalls wherever a human has to press power and read serial, so this build puts an Orange Pi in that seat and gives the agent one flash command that boots over USB and writes the eMMC.
Reality
Patrick's Context Development Lifecycle treats skills and rules files as software with four phases. The phase teams skip is evaluation, the only one that catches a model upgrade changing behaviour while the file stays byte-identical.
Reality
The leverage in this rebase sits in routing rather than prompting, and the three regexes that decide which hunks skip the model entirely are tuned to one repository's five months of noise. Copying the method means re-deriving them.
Reality
The AI Frontier Division built Cofa-Probe on OpenAI's Codex to score four stages of human-agent collaboration, which tells you roughly what Krafton now thinks a finished portfolio piece is worth as hiring evidence.
Reality
The new Code Repository and AI Code Review are built for thousands of simultaneous pull requests, and the design concedes that the slow part of shipping agent-written code was always waiting for a human to look at it.
Reality
JetBrains has pointed the next four months of CLion at embedded and agentic debugging, with the debugger config itself becoming a JSON file you can edit by hand, hand to an agent, or import from VS Code.
Reality
Block has published the first hard distribution of AI adoption inside a large engineering org: one agent for most, three to five for the next cohort, and a small team building the orchestrator.
Reality
Munder Difflin's founder says a Claude Code agent redrew a whole blog at no API cost. The reusable SVG system, the no-red rule and the approval gate are the parts that transfer.
Reality
A dev.to writeup credits Google Jules with a signed Android release produced from a phone browser. What stood in for the laptop was a GitHub Actions runner holding the SDK and the keystore.
Reality
A developer reports driving a CASE repository of 24,483 catalogued elements from sqlplus alone: 722 CI_* views to read, about 200 CIO* packages to write, and no surviving documentation.
Reality
Version 1.0.0 landed as one signed commit into an empty repository, with CI-built binaries. When agents write most of the code, the release becomes the only boundary a reviewer can hold.
Reality
A vendor walkthrough has a coding agent build an endoscopic segmentation app from HoloHub examples, with the CLI as the shared execution surface and named skills as the spec.
Reality
A New Stack column argues AI-generated diffs have made careful review impossible. The automatable half of the job is bug-catching; the half nobody has staffed is knowledge transfer.
Reality
Build · August 19, 2026 · 1 publisher
Build · August 19, 2026 · 1 publisher
Build · August 19, 2026 · 1 publisher
Build · August 17, 2026 · 1 publisher
Build · August 15, 2026 · 1 publisher