A collection of Claude Code plugins for productivity, learning, and software-development workflows.
# Add this marketplace
/plugin marketplace add joshuaoliphant/claude-plugins
# Install any plugin
/plugin install <plugin-name>@oliphant-pluginsNine plugins, grouped by focus. Current versions live in
.claude-plugin/marketplace.json and each plugin's
plugin.json (the source of truth) — they are intentionally not duplicated here so this
list can't drift.
Create evidence-based spaced-repetition flashcards using cognitive-science principles (Andy Matuschak's five properties of effective prompts). Quality validation, knowledge-type workflows (factual, conceptual, procedural, salience), anti-pattern detection, and deck/card management via the Mochi API.
Requires: a Mochi account + MOCHI_API_KEY.
/plugin install mochi-creator@oliphant-plugins
export MOCHI_API_KEY="your_api_key_here"Institutional memory across sessions and projects: capture solved problems, retrieve past solutions before you start, and graduate recurring lessons into living context (CLAUDE.md / AGENTS.md). Structured YAML-frontmatter files for fast grep-based retrieval.
Skills: compound-capture, compound-retrieve, compound-graduate.
Process information for real understanding — an antidote to the illusion of clarity. The
explain-back skill makes you explain from memory, grades you against the real source,
teaches struggle-then-teach per gap, and outputs Mochi cards plus a resumable session
record.
Autonomous SDLC as a state machine on disk, driven by a Stop-hook loop (optionally a
user-armed /goal) — one verified, committed unit of work per iteration with
decide-log-proceed autonomy and built-in verify / code-review / simplify gates. Architect
and Builder subagents; TDD, BDD, and Beads skills.
Build web apps where an AI agent dynamically generates the HTML UI, using hexagonal / ports-and-adapters architecture with HTMX for interactivity and MCP tools for data operations.
Stack: Claude Agent SDK + FastAPI + HTMX + Tailwind CSS. Requires: ANTHROPIC_API_KEY.
/plugin install hexagonal-agents@oliphant-plugins
export ANTHROPIC_API_KEY="your_key_here"Generate autonomous experiment loops (Karpathy-style autoresearch) that edit → run →
measure a scalar metric → keep improvements via git. Produces program.md plus an
immutable auto/run.sh with tiered quality gates and structured METRIC output, ready to
run with claude --dangerously-skip-permissions.
Install a local, Docker-free OTLP observability stack (OpenTelemetry → Vector → JSONL,
optionally VictoriaLogs + VictoriaMetrics) into a project and instrument the code — with a
scan-and-propose step for domain-appropriate spans and metrics — then query that telemetry
as a development feedback signal (observability-query skill, status.sh --json detection
contract, scripted end-to-end verify.sh). Composed by autonomous-sdlc as a soft
dependency.
Local web-based diff review: comment on your working-tree or committed-branch diff in the browser and feed the comments back to Claude Code.
A manually-driven ("disassembled") SDLC — the same .sdlc/ session format as
autonomous-sdlc, but you drive each phase by hand via slash commands
(/spec → /plan → /build → /verify → /journal). Built for legible, narratable live demos;
adopts an in-progress autonomous-sdlc session and stands its loop down.
Each plugin owns its version in plugins/{name}/.claude-plugin/plugin.json; that version is
copied into .claude-plugin/marketplace.json at publication time. Run
python scripts/check_marketplace_versions.py to confirm they match. See
CLAUDE.md for full development and versioning guidance.
Contributions are welcome — feel free to submit issues or pull requests.
MIT