task-observer: docs/ai/writing-docs updates from 2026-05-25#32
Open
divad12 wants to merge 1 commit into
Open
Conversation
Captures four observations about preserving enforcement strength when tightening agent instructions: imperative wording, rationale for judgment-heavy rules, failure-mode explanation for safety-critical rules, and positive framing for tradeoff guidance.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Target\nglobal-doc: docs/ai/writing-docs.md\n\n## Observations addressed\n- Observation 1 (dotfiles/log.md): Compression weakened corrective guardrails → changed: added ## Compression section; rule to preserve imperative wording that counters known failure modes even when cutting text\n- Observation 2 (dotfiles/log.md): Explain why guardrails exist for judgment-heavy rules → changed: added bullet to keep a short rationale sentence for judgment-heavy rules so the model generalizes correctly\n- Observation 3 (dotfiles/log.md): Code guardrails need failure-mode rationale → changed: added bullet to keep failure-mode explanation (what goes wrong, stop signal, allowed action) for safety-critical rules\n- Observation 4 (dotfiles/log.md): Positive phrasing is better for tradeoff guidance → changed: added bullet to lead with positive framing for tradeoff rules instead of prohibitions
Generated by Claude Code