New Skill Audits Projects for Legacy Instructions Written for Older Claude Models
daniel_mac8 · x · 2026-11-11
Developer danielmac8 released an installable skill, /fable-5-2-prompt-audit, based on Anthropic's official guide section on revisiting instructions written for earlier models. It scans projects and flags seven categories: standing subagent delegation, hardcoded token budgets, harness code that edits earlier turns (Fable 5.2 uses 'preserved thinking', so history should be append-only), 'treat that answer as done' lines, instructions to think less, rigid prose style limits, and stale cautious rules piled up in rules/notes/memory. It also verifies its checklist against the live guide and lists every place a project stores instructions (instruction files, .claude/, skills, memory).
Related event: Developer Open-Sources Fable 5.2 Prompt Audit Skill(2 posts)→
More from coding & agent
- Meme: what your subagents do when the task needs 5 lines of code — neal_lathia · 2026-11-11
- Visa tested 22 models on 15 shopping tasks: ~43% failure, agents bought wrong items anyway — Thick_Mulberry_2235 · 2026-11-11
- MCP or CLI? Yes — Sentry ships a unified experience bridging both — zeeg · 2026-11-11
- Stanford Launches CS329Z: Engineering AI Agents, Covering Memory, Evals and Proactive Agents — stanfordnlp · 2026-11-11
- Solo dev took down his multi-agent site for 8 hours by draining the shared model budget — wolfdogKC · 2026-11-11
- Stanford CS329Z releases assignment on building terminal-bench-style agent evals with Harbor — Diyi_Yang · 2026-11-11