Skip to content

task-observer: docs/ai/writing-docs updates from 2026-05-18#23

Open
divad12 wants to merge 1 commit into
masterfrom
task-observer/writing-docs-2026-05-18
Open

task-observer: docs/ai/writing-docs updates from 2026-05-18#23
divad12 wants to merge 1 commit into
masterfrom
task-observer/writing-docs-2026-05-18

Conversation

@divad12
Copy link
Copy Markdown
Owner

@divad12 divad12 commented May 18, 2026

Target\nglobal-doc: docs/ai/writing-docs.md\n\n## Observations addressed\n- Observation 1 (dotfiles/log.md): Compression can weaken corrective guardrails → changed: added rule to preserve enforcement strength (wording that counters known failure modes) even when cutting for context budget\n- Observation 2 (dotfiles/log.md): Explain why guardrails exist when model behavior depends on it → changed: added rule to keep rationale for judgment-heavy rules, since stripping the "why" turns a calibrated rule into a bare imperative\n- Observation 3 (dotfiles/log.md): Code guardrails need failure-mode rationale → changed: added rule to preserve failure-mode context (what goes wrong, stop-signal thought pattern, safe alternative) for hard rules\n- Observation 4 (dotfiles/log.md): Positive phrasing is better for tradeoff guidance → changed: added rule to prefer positive phrasing when the correct alternative is simple and actionable\n\nAll four observations target the same compression scenario and are grouped into a new "When Tightening Existing Instructions" section in writing-docs.md, placed after the general Contract bullet list.


Generated by Claude Code

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant