prompt

The same agent task produced different results every run. The prompt left decisions to the model; the fix is an explicit work specification.

TL;DR: Same prompt, different outputs usually means the prompt underdefined the work, not that the model is unstable. Treat prompts as work specifications: state objective, context, constraints, output shape, and a verification step. Clear specs make runs converge, checking objective, and cut the expensive corrective iterations, regardless of which model you use.

A 228-line prompt that fought itself became a 205-line shared contract plus four swappable mode files. The design logic, explained.

TL;DR: Splitting one bloated 228-line prompt solved conflicting instructions by separating rule ownership: a shared contract locks universal invariants like citation and privacy rules, while swappable mode files handle tone and structure. Four modes—report, news, tutorial, explainer—are chosen by the promise made to readers. Hard structure limits and strict evidence requirements keep every output consistent and checkable.

Briefs headed to an external chatbot still carried hostnames and real repo names. A three-checkpoint mechanical gateway shut it.

TL;DR: Internal data was leaking to an external LLM, an architecture flaw not a prompt issue. The fix is a three-stage gateway that rewrites real identities to fictional ones before the model sees them, blocks pushes with real tokens, and scans published posts. It uses reserved example domains and doc IPs, and found seven leaks in 564 articles.

The transcript must carry the hijack sentence verbatim as data. That is exactly how you prove a media file never acted as an instruction.

TL;DR: The author deliberately embedded a prompt-injection sentence into an audio fixture to test a media pipeline's defenses. Counterintuitively, the test requires the hostile transcript to stay verbatim and labeled untrusted, since injection is about text switching from data to instructions. The real defense is a contract proving content never becomes commands, not filtering text or trusting the model's intelligence.

Machine prompts on my cron pipeline moved to English for one reason: one control variable across a router of very different models.

Last week, my blog article pipeline failed completely. I thought pasting a 100KB skill containing hundreds of style rules into the prompt would solve the...

A repo with many coding agents needs more than one instruction file: AGENTS.md for procedures, user memory for preferences, a fact store for on-demand facts.

TL;DR: Repeated agent mistakes in a repo pointed to missing documentation, not a bad agent. The fix is a three-layer setup: AGENTS.md for repo procedures, user memory for personal preferences, and a fact store for reference material. Claude Code reads the same rules through a CLAUDE.md that imports AGENTS.md, keeping every session consistent.

A raw CTA_FLAG: YES marker leaked into a live chat bubble. The strip regex only matched end-of-string, and the SSE stream bypassed it entirely.

TL;DR: A CTA_FLAG marker leaked into the UI because the model ignored placement instructions and SSE streaming sent chunks before the strip regex ran. The fix moved CTA data into a structured field on the knowledge source, rendered directly by the frontend. Lesson: never parse UI-critical data out of model prose; use structured fields.

A RINGKAS label leaked into the chat bubble. The fix was not a stricter prompt but sanitizing input and output at both ends of the pipeline.

TL;DR: The author's chatbot kept leaking prompt labels like RINGKAS into answers, and adding prompt prohibitions never fixed it. They moved control to code: sanitizing knowledge-base text before injection and rendering output through react-markdown with rehype-sanitize allowlists. Lesson: prompts give rough direction only; reliable formatting enforcement belongs at input and output boundaries.