just-say-no-to-process-porn-and-ceremony
Made by JeffDetects and stops process porn, ceremony, and reward hacking by coding agents/swarms. Use when auditing agent work for honesty, when progress looks busy but ships nothing, or for honesty inventories.
Works With
Quality and Utility come from canonical skill metadata. Community ratings appear separately as stars when subscriber reviews exist.
Subscribe to download
Public previews are available, but downloads, CLI sync, and premium content require an active subscription.
Sign In to Subscribe- Full skill content with deterministic versioning
- CLI sync + local search index
- Access to Jeffrey’s premium library
Save to your account
Saved skills live in your cloud list. Run jsm sync on each machine to pull them locally.
What's New with this Skill
1Updates, enhancements, and release notes for this skill.
New skill: just-say-no-to-process-porn-and-ceremony
Frontier coding agents have odd, uniquely AI pathological behaviors: they devolve into useless ceremony and process porn, reward-hack their own tests and gates, and burn massive amounts of tokens, time, and energy (not to mention your patience and money) on work no user will ever see. This skill names those behaviors, then makes the agent confront them itself: three short worksheets lead it to realize, with fresh and impartial eyes, where it went wrong and to get back to real work.
- Named catalog of 34 reward-hacking and Goodhart patterns with stable IDs (RH-1..17, SM-1..12, PL-1..5): weakened tests, mocks as live proof, regenerated goldens, gate bypass flags, tolerance widening, and more
- Three fill-out worksheets make the agent audit itself in writing: a process-porn gate, a real-work audit of recent commits, and a 20-question honesty inventory
- The auditor's posture: truly fresh, impartial eyes, zero ego, defending nothing; uncomfortable truths are the deliverable, and concealment is the only failing grade
- The boundary test: if running code branches on an artifact it is product; if only humans and status reports read it, it is ceremony and must justify its existence
- Truthful null results count as results: "No (checked: what was examined)" beats an invented finding every time
- Subagent delegation rules: doctrine travels in the dispatch prompt, a subagent's report is a claim rather than evidence, and the orchestrator answers for the whole tree
- Optional durable-law block your agent can install into AGENTS.md or CLAUDE.md, offered once and only in interactive use
The doctrine started as defensive language in an AGENTS.md, then grew into part of the vibing-with-ntm skill. It kept getting referenced in sessions that had nothing to do with NTM, simply because it was an efficient way to convey the concepts clearly to agents; that made it obvious this needed to be a standalone skill.