PaulRBerg.

agents-context-management

Create or polish repo agent context: README.md, AGENTS.md/CLAUDE.md, and installed project skills.

agents-introspection

Determine whether prior Codex and Claude Code work in the current project establishes a recurrence risk for the user's task, then recommend the smallest durable intervention justified by the evidence.

audit-stale-comments

Audit source comments against the code they describe. Report only confirmed mismatches and noise; do not edit files unless the user separately asks for fixes.

bump-deps

Use for dependency updates: bump npm/pnpm/yarn/bun packages, check outdated, or run taze.

bump-release

Cut a release: bump versions, write changelogs, commit, tag.

claude-handoff

Orchestrate one to five Sonnet subagents to implement an approved Claude Code plan.

cli-cast

Separate read, preparation, simulation, signing, and broadcast so no state-changing action is hidden inside command construction.

cli-gh

Route GitHub CLI work through current `gh` help and load only the reference for the active task.

cli-just

Create readable task automation that matches the repository's installed Just version and existing justfile conventions.

code-polish

Resolve scope once, make only high-confidence simplifications, fix evidenced defects by risk, and verify the final state.

codex-handoff

Orchestrate one to five Codex CLI agents to implement an approved Claude Code plan.

coingecko-cli

Use the installed CLI's machine-readable command catalog as the source of truth for supported market-data operations.

coingecko-open-page

Open the CoinGecko historical-data page for a coin/date in Chromium via Chrome DevTools MCP.

effect-ts

Apply Effect semantics using local project patterns first, then only the task-specific reference and upstream source needed to resolve uncertainty.

evm-atlas

Resolve and query only the target mainnets in `references/generated/target-mainnets.json`, under a strict read-only boundary.

find-tool

Recommend the current option that best fits the user's actual constraints, backed by fresh primary evidence.

fresh-eyes-sweep

Inspect the requested Git scope for evidenced mistakes, fix every safe issue, and continue until every mapped file is accounted for and a full pass finds nothing new. A verified no-op is valid only after that coverage.

git-squash

Squash a feature branch into one commit via soft reset to the merge base, ready for a clean PR.

html-debrief

Use for interactive HTML debriefs or saved HTML findings/reports from the current task.

html-playground

Build a self-contained interactive HTML explorer with controls, live preview, and a copyable natural-language prompt.

large-file-refactor

Use LOC thresholds to discover candidates, then decide whether a split is justified by cohesion, coupling, ownership, and change risk. Test files use a relaxed 2000 LOC discovery threshold.

naming-refactor

Refactor naming and repository structure exhaustively while preserving behavior and external contracts.

night-shift

Autonomous overnight codebase improvement with bounded runtime, evidence-gated changes, and verification.

repo-rename

Preview the complete GitHub, Git, filesystem, and agent-continuity mutation set before applying a repository rename.

skill-map

Find skill installs and references across local files without scanning macOS protected home paths or obvious transcript, cache, dependency, and backup noise.

spreadsheets

Handle tabular data with exact values, minimal diffs, local privacy, atomic writes, and structural validation.

tailwind-css

Style within the installed Tailwind version, existing design tokens, component patterns, and product context.

todo-archive

Archive checked TODO.md tasks into `.ai/todos/TODO_UNTIL_YYYY_MM_DD.md`, leaving unchecked tasks.

yeet

Create or update GitHub contributions from repository evidence, using the matching workflow's templates, idempotency rules, and Paul's writing voice.