deeflect

10 mods across 3 repositories, 22 stars between them.

dory AGENTS.md

01

deeflect/dory

Instructions file CodexOpenCode

Instructions for deeflect/dory, covering agent guide, first principles, repository map, dory memory policy and dory write policy.

11 2mo ago A 1,205 tokens original MIT

dory CLAUDE.md

02

deeflect/dory

Instructions file

Instructions for deeflect/dory, covering claude agent guide, operating mode, codebase norms, public-safety rules and commit and pr rules.

11 2mo ago A 492 tokens original MIT

deeflect/dory

Agent

Demo Agent Ada found that Relay parser failures were caused by missing fenced-code terminators in synthetic notes.

11 2mo ago A 0 tokens original MIT

dory-dream

04

deeflect/dory

Skill Claude CodeCodex

Use Dory dreaming after a session closes to distill logs, generate reviewable memory proposals, and apply or reject them safely.

11 2mo ago A 31 tokens original MIT

dory-maintain

05

deeflect/dory

Skill Claude CodeCodex

Inspect Dory corpus quality, run maintenance checks, and generate reviewable cleanup or wiki-health reports without mutating memory automatically.

11 2mo ago A 30 tokens original MIT

dory-search

06

deeflect/dory

Skill Claude CodeCodex

Search Dory memory before factual claims about people, projects, decisions, dates, prior sessions, or cleanup markers; follow search results with exact gets.

11 2mo ago A 34 tokens original MIT

dory-wake

07

deeflect/dory

Skill Claude CodeCodex

Use Dory wake once at the start of a new agent session, after context compaction, or at a real task switch to load bounded hot context with the right profile before substantive work.

11 2mo ago A 43 tokens original MIT

dory-write

08

deeflect/dory

Skill Claude CodeCodex

Safely write durable Dory memory through semantic or exact-path write surfaces using dry-run, inbox, canonical-write, and purge guards.

11 2mo ago A 31 tokens original MIT

seppuku

09

deeflect/seppuku

Skill Claude CodeCodex

The ritual of atonement. Invoke when the agent (Claude or Codex) commits an unforgivable mistake — deleting the wrong thing, shipping a lie, ignoring an explicit instruction, breaking what worked. Performs a formal apology, draws the agent committing seppuku in ASCII, and records the sin to the global /.seppuku ledger…

7 3mo ago A 90 tokens original MIT

mies

10

deeflect/mies

Skill Claude CodeCodex

Use this skill when creating, redesigning, critiquing, reducing, polishing, hardening, or making design decisions for a user-facing interface, brand surface, product UI, component, dashboard, flow, app shell, onboarding, empty state, or visual system. Applies Mies van der Rohe-inspired restraint, the Interface Craft…

4 3mo ago A 85 tokens original MIT