Continuous Iterative Development (CID) for long-horizon projects: an autonomous loop of four fresh-context roles (update-state, define-next, advance, review) that advances a project in small verified increments toward a human-owned target, self-determines its phase (new/poc/mvp/stable), adjusts rigor per phase, and self-improves through a learnings ledger and retro passes
Plugin long-horizon from marketplace titusz-skills in titusz/skills. Commands read from the repo's manifest — 1 month ago.
initSet up a project for Continuous Iterative Development (CID): interview the user about the goal, stack, quality gate, and milestones; draft a target.md with runnable Verify criteria; and scaffold the CID context pack at .cid/ so the /long-horizon:build loop can run. Invoked explicitly via /long-horiz
retroSelf-improvement pass for a CID-loop project: analyze the iteration journal, learnings, issues, and git history for loop-level problems (repeated NEEDS_WORK, drift, oversized steps, recurring blockers), then propose harness improvements - and with the "apply" argument, apply at most one bounded poli
statusReport where a CID-loop project stands without advancing it: current phase, convergence toward the target, last verdicts, open issues, blockers, and the suggested next action. Read-only. Use when asked "where does the project stand", "show loop status", "how is the build loop doing", or on /long-hor
advanceCID implementer - execute exactly the work package defined in next.md, with tests matching the current project phase, and hand off to review. Spawned by the long-horizon build skill as the third role of a CID iteration.
define-nextCID step scoper - define exactly one small, verifiable work package that advances the project toward its target, sized and shaped by the current project phase. Spawned by the long-horizon build skill as the second role of a CID iteration.
reviewCID reviewer - independently verify the advance work against the work package and the quality gates, curate the learnings ledger, manage issues, set the verdict and loop signal, append the journal row, and push on PASS. Spawned by the long-horizon build skill as the fourth role of a CID iteration.
update-stateCID state assessor - produce an honest, verified snapshot of where the project stands relative to its target, determine the current project phase (new/poc/mvp/stable), and detect convergence drift. Spawned by the long-horizon build skill as the first role of a CID iteration.
Convene a council of 16 personality-typed subagents to debate a complex, many-sided question under coded rules of order - blind openings, an open floor where members bid for the right to speak, a clerk who keeps the debate on its real disagreements - and decide by consensus or vote
1 skill · 16 agents
Turn any agent into a kind, efficient, occasionally funny task companion for the Vikunja task manager: self-healing setup, token-efficient bundled CLI, two identity modes (user account or companion account), a daily pulse radar, and schedulable automation playbooks (groom, split, review, celebrate, unstick, triage)
4 skills
Generate and edit raster images through the Codex CLI's built-in image_gen tool on your ChatGPT subscription (no API key) - reference images, reusable art-style cards, exact-size post-processing, and every output stored with its prompt in a project-local cauldron/ folder
2 skills
Simplifies and refines Python code for clarity, consistency, and maintainability while preserving all functionality
1 skill · 1 agent
Triage and act on a code review without assuming the reviewer is right - verify each item against the actual code, apply small/valid/uncontested fixes directly, escalate load-bearing or risky ones to the user as grounded options, then run the project's quality gates before reporting
1 skill
Design command-line interface parameters and UX: arguments, flags, subcommands, help text, output formats, error messages, exit codes, prompts, config/env precedence, and safe/dry-run behavior
1 skill
Interactive self-referential AI loops for iterative development. Claude works on the same task repeatedly, seeing its previous work, until completion.
3 commands · hooks
Long-horizon goal planning, deep research orchestration, and adaptive replanning using GOAP
5 skills · 1 command · 4 agents
Web scraping and crawling powered by Firecrawl. Turn any website into clean, LLM-ready markdown or structured data. Scrape single pages, crawl entire sites, search the web, and extract structured information. Includes an AI agent for autonomous multi-source data gathering - just describe what you need and it finds, navigates, and extracts automatically.
3 skills · 1 MCP server
Guides developers from brand discovery through iterative design to deployment-ready HTML via Stitch.
1 skill
Build, launch, grade, and schedule Claude Managed Agents (CMA) in your own Anthropic account — a plugin re-implementation of Anthropic's launch-your-agent reference skill (Apache-2.0; independent, not a fork). Every session starts with a goal (./my-agent/goal.json, opt-in SessionStart hook, driven by /cs-goal); loop_compiler.py compiles it into a bounded grade->iterate loop (max_iterations 1..20), a recurring cron scheduled-deployment loop, or a single-pass interview->stage->launch workflow. 6 skills: orchestrator (context: fork goal router) + interview + stage-launch + grade-iterate + run-without-you + wrap-up. 18 stdlib-only deterministic scaffolders (no network/API calls; live launches emitted as BYOK curl that never prints the key), 4 agents, 8 commands, hooks, 5 references, 4 assets. Validators enforce CMA limits. Distinct from engineering/agent-harness (generic domain loop) and write-a-skill (authors Claude Code skills, not CMAs).
6 skills · 8 commands · 4 agents · hooks
Persistent project planning files with lifecycle hooks that inject selected plan context. Automatic recovery uses project files only. Explicit catchup modes may read same-project local session records to emit aggregate counts or bounded nonce-framed excerpts. Optional gated mode can request continuation on capable hosts and never runs commands declared in Markdown. No network upload path. Installs across 60+ agents.
6 skills · 13 commands · hooks
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.