Why I Named a Governance Rule 'Cipher Pol'

A friend opened the Stella Protocol codebase and asked: why do your skills have names like secret intelligence agencies?

Short answer: because that is exactly what I was going for.

The naming map

Stella Protocol — my open-source AI-PM methodology — is built on top of One Piece references. Not for cute. Every name has to invoke its function in my head without me having to open the docs.

• Cipher Pol = scope drift monitor. In One Piece, Cipher Pol is the government’s intelligence agency — their job is to sniff out suspicious activity. Perfect fit for a skill that checks “hey, this new route, is it in the PRD?”
• Buster Call = quality/security veto. A Buster Call is the highest-level emergency strike. If there’s a hardcoded secret or missing auth, this skill blocks the phase transition. Fitting name for something with stop-the-line authority.
• Vivre Cards = append-only decision log. A Vivre Card is a piece of paper pointing to someone that cannot be erased. Append-only, immutable, always pointing to something.
• Punk Records = project brain files. Punk Records is Vegapunk’s research archive. The place where all project context lives: log-pose.md, architecture.md, vivre-cards.md.
• Observation Haki = 7-lens PRD framework. Haki is a perception ability. A good PRD sees things before they become problems — user need, edge case, dependency — so the name fits.
• Log Pose = project phase tracker. A Log Pose is a compass that points to the next island. log-pose.md stores current phase + next gate.
• Grand Line vs East Blue = track ceremony level. Grand Line is the dangerous main route that needs preparation. East Blue is the calm sea. Complex features take the Grand Line track (full ceremony: PRD, architecture, design, review). Simple features take East Blue (lightweight).
• Shaka, Pythagoras, ODA, Lilith = Vegapunk satellites = specialized agent skills. Shaka writes PRDs. Pythagoras handles architecture. ODA does design. Lilith reviews. One satellite, one domain.

Why evocative beats descriptive

“scope-drift-monitor” is informative. But it does not stick. A week later I forget the exact name — scope-monitor? drift-checker? scope-change-watchdog? I have to open the repo to remember.

“Cipher Pol” sticks on first encounter. Because it is not just a label, it comes with a narrative association: intelligence agency. My brain already has a folder for “intelligence agency” long before I built this skill. I just park the skill’s function in a folder that already exists.

This is not abstract naming theory. It is memory retrieval. Names with narrative hooks get recalled faster than names that are just function descriptions.

The three rules I use

1. Evocative beats descriptive. A name should invoke an image or a feeling, not just describe function. “Buster Call” invokes urgency. “Quality Gate Enforcement Mechanism” invokes sleep.
2. Culturally anchored for the author. I live in One Piece — I know Cipher Pol without thinking. You might live in Star Wars, Harry Potter, or a favorite football club. Pick what you have already internalized. Don’t borrow a reference you yourself have to Google.
3. One word beats a phrase. “Cipher Pol” vs “Cross-Reference Scope Monitoring Engine.” Two syllables vs six. The first fits in a commit message, a skill file header, and spoken conversation. The second only fits in a slide.

The caveat: this is naming for the author

Stella Protocol is a solo project. I am the only user who has to remember every name. So the One Piece density is fine — the associations work for me.

If your project is shared with 10 people who don’t all read One Piece, the naming has to sit on common ground. That doesn’t mean generic — it means the anchor has to be something the team actually shares. A team of engineers who all watched Breaking Bad can use “Heisenberg” for override mode. A team that didn’t, can’t.

The rule is not “use pop culture.” The rule is: pick an anchor your naming audience is already familiar with.

An example that did not work

An early version of Stella had a skill named “Shirohige” for archiving (Whitebeard — old captain, big presence, symbol of legacy). My logic: archive = legacy = Whitebeard.

The problem: Shirohige is a character, not a place. When I said “this log goes to Shirohige,” my brain got confused — goes to a character? So I switched to Punk Records — Vegapunk’s literal archive. Clear place, clear action (“save it to Punk Records”).

Lesson: even within the same reference set, not every name is equal. The winner is the one whose grammar matches how you’ll actually talk about the function.

Lesson

Naming is UX for the developer’s brain, not ornament. If you have to open the docs to remember what something does, the name failed.

Related: Append-Only Decision Log — why Vivre Cards have to be immutable.

Key Takeaways

1. Narrative hooks beat descriptive labels. “Cipher Pol” is retrieved faster than “scope-drift-monitor” because the brain already has a folder for “intelligence agency.” Park the function in the folder that exists.
2. Anchor the reference in shared ground, not private ground. Solo project? Use the fiction you’ve internalized. Shared team? The anchor has to be something the team actually shares. Otherwise you’ve built a private joke.
3. Grammar matters as much as reference. “Save it to Punk Records” works; “save it to Shirohige” does not. The name has to fit the sentence you’ll speak or commit, not just the theme.

Satellite: Morgans (this post) · Pipeline: REFLECT — Morgans