By dkod-io
Git-native flight recorder for AI coding agents — capture every session into your repo's git refs; blame any line back to its prompt; flag intent drift
Trace lines of a file back to the AI session and prompt that produced them (dkod blame <path>)
Flag intent-vs-output drift in captured AI sessions (dkod drift)
Initialize dkod session capture in the current repository (dkod init)
Show captured dkod sessions for this repo (dkod log)
Matches all tools
Hooks run on every tool call, not just specific ones
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
No model invocation
Executes directly as bash, bypassing the AI model
No model invocation
Executes directly as bash, bypassing the AI model
Git-native flight recorder for AI coding agents. This plugin wires dkod into Claude Code so every session is captured into your repo's own git refs (refs/dkod/sessions/*) — no external service, no SaaS database. Blame any line back to the prompt that produced it. Flag intent-vs-output drift.
Privacy: Transcripts are never stored outside your git host.
SessionStart, UserPromptSubmit, PreToolUse, PostToolUse, PostToolUseFailure, PreCompact, Stop, SessionEnd) stream the session into refs/dkod/sessions/<id> in the current repo.dkod binary is missing or misbehaves, every hook still exits 0 and your session is never blocked./dkod:init — initialize capture in the current repo (dkod init)/dkod:log — list captured sessions (dkod log)/dkod:blame <path> — trace lines back to the session/prompt that wrote them (dkod blame)/dkod:drift — flag intent-vs-output drift (dkod drift)The dkod CLI (v0.2.0+) must be installed:
brew install dkod-io/tap/dkod
or:
curl -fsSL https://raw.githubusercontent.com/dkod-io/dkod-cli/main/install.sh | sh
From this repo's marketplace:
/plugin marketplace add dkod-io/claude-plugin
/plugin install dkod@dkod
Or load directly from a local clone for development:
git clone https://github.com/dkod-io/claude-plugin
claude --plugin-dir ./claude-plugin
Then run /dkod:init in any repo you want captured.
Each hook runs:
command -v dkod >/dev/null 2>&1 && dkod capture-hook --agent claude-code --event <Event> || true
dkod capture-hook buffers the hook event and exits 0 on misuse by design; the command -v guard and || true make the hook a no-op when dkod isn't installed or a stale binary errors — your session is never interrupted (see dkod-io/dkod-cli#24).
Transcripts are never stored outside your git host. Sessions live as ordinary git objects under refs/dkod/sessions/* in your repository — you control where they go, who reads them, and when they're deleted, exactly like any other ref.
MIT
npx claudepluginhub dkod-io/claude-plugin --plugin dkodLocal-first parallel AI coding agents. N Task subagents rewrite different functions in the same file simultaneously; AST-level merge happens in-place via the dkod-mcp server. One worktree, one dk-branch, one PR. Zero network coordination.
Continuous self-referential AI loops for interactive iterative development, implementing the Ralph Wiggum technique. Run Claude in a while-true loop with the same prompt until task completion.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.
Complete Claude Code configs for mobile development - Android, iOS, and Kotlin Multiplatform
Consult multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, plus codex and gemini CLIs when installed) to get diverse perspectives on coding problems
Advanced development superpowers for orchestrating complex workflows with Superpower Loop and work verification