randroid-loop
/randroid-loop
Run a bounded research or implementation loop with persistent Dots task tracking and an explicitly selected git workflow. Use when the user asks for an autonomous multi-iteration development loop, research loop, implementation loop, or work-until-complete cycle. Do not use for an ordinary one-pass coding task.
- Author
- randroids-dojo ↗
- Repository
- skills ↗
- Licence
- MIT
- Last repository change
- 07.10.2026
Assessment and security
- Files in this package
- 9
- Scripts
- 2
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 07.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
- 2 scripts in this package. Review them before installing, as Claude may execute them.
Scripts: scripts/randroid-loop.sh, scripts/setup-loop.sh
What it does and when to use it
Run a bounded research or implementation loop with persistent Dots task tracking and an explicitly selected git workflow. Use when the user asks for an autonomous multi-iteration development loop, research loop, implementation loop, or work-until-complete cycle. Do not use for an ordinary one-pass coding task.
Original author description
Run a bounded research or implementation loop with persistent Dots task tracking and an explicitly selected git workflow. Use when the user asks for an autonomous multi-iteration development loop, research loop, implementation loop, or work-until-complete cycle. Do not use for an ordinary one-pass coding task.
How it works
- Before starting
- Select the task tracker
- Run the loop
- Bundled helpers
SKILL.md preview
--- name: randroid-loop description: Run a bounded research or implementation loop with persistent Dots task tracking and an explicitly selected git workflow. Use when the user asks for an autonomous multi-iteration development loop, research loop, implementation loop, or work-until-complete cycle. Do not use for an ordinary one-pass coding task. --- # Randroid Loop Run a sequence of independently verifiable work iterations whose durable state lives in the target repository. ## Before starting 1. Read the target repository's agent instructions and current task tracker. 2. Ask for any missing choices that cannot be inferred from the user's request: - mode: `research` or `implement`; - stopping condition: an exact iteration count, until complete, or explicitly requested continuous operation; - git workflow: no commit, commit only, push, open PR, or PR and merge; - whether each iteration should use fresh context; - which Dots tracker to initialize when the repository has no existing Dots state; - any run-specific priorities. 3. Treat push, PR creation, merge, deployment, publication, and other external mutations as separate authorization boundaries. Never select them merely because the repository appears ready. 4. Require a finite iteration count or `until complete` unless the user explicitly requests a continuous loop. ## Select the task tracker - Use `dot-html` when `.dots/` contains HTML tasks and the command is available. - Otherwise use `dot` when `.dots/` contains Markdown tasks and the command is available. - Do not mix trackers in one project. - If no Dots state exists and both CLIs are available, ask which format to initialize. If neither tracker is available, stop and report the missing prerequisite instead of inventing replacement state. ## Run the loop 1. Read [references/research-loop.md](references/research-loop.md) only in research mode, or [references/implement-loop.md](references/implement-loop.md) only in implementation mode. 2. Read [references/loop-shared.md](references/loop-shared.md) for the shared task, verification, git, and completion contract. 3. In each iteration: - re-read current repository and task state; - select one ready, bounded item; - perform only the selected mode's work; - run proportionate verification; - update the task and repository ledgers; - perform only the authorized git workflow; - emit a concise iteration receipt. 4. Stop when the requested count is reached, the shared completion contract is satisfied, user input is required, or an authorization boundary is reached. 5. Do not report completion while required verification, remote state, or an authorized delivery step remains unresolved. ## Bundled helpers Resolve these paths relative to the directory containing this `SKILL.md`: - `scripts/setup-loop.sh` writes local loop state. Run it only after the choices above are settled. - `scripts/randroid-loop.sh` is an optional fresh-context Codex CLI wrapper. Review its selected mode, stopping condition, and git workflow before launching it. - [references/looping-design.md](references/looping-design.md) explains the architecture; read it only when modifying or diagnosing the loop itself.
Files in this package (9)
- SKILL.md 3 KB
- agents/openai.yaml 1 KB
- references/implement-loop.md 7 KB
- references/loop-shared.md 4 KB
- references/looping-design.md 3 KB
- references/research-loop.md 6 KB
- scripts/randroid-loop.sh 13 KB
- scripts/setup-loop.sh 6 KB
- state/.gitignore 1 KB
Versions
The database has no other skills named randroid-loop.
Database snapshot: 13.11.2026 · package version 3a089d14559b (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 86randroid
Route requests to the focused Randroid development workflows and explain which capability to use. Use when the user asks what Randroid skills are available, invokes the Randroid namespace without naming a workflow, or needs help choosing among loop execution, PR feedback… - 95ralph-loop
Execute an autonomous development loop that picks one task per iteration, implements it, verifies it, and commits the result — each iteration in a fresh context window. Use when user runs /ralph, mentions "ralph loop", "autonomous loop", "builder verifier", "run tasks… - 100loop-engineering
Use when a repeatable task must become a bounded Trigger -> Execute -> Verify -> State loop, scheduled automation, goal agent, or metric-driven research cycle. - 100ralph
This skill should be used when the user wants to set up an autonomous coding loop, plan a Ralph loop, prepare for headless Claude execution, create a spec and implementation plan for autonomous coding, or run an autonomous development workflow. Trigger phrases include "ralph",… - 96loop
Use only when user explicitly invokes $loop. - 100loop-designer
Compile a goal into a durable, file-backed agent loop — a .loop/<name>/ charter any coding agent (Claude Code, Codex, Gemini, Grok) can execute unattended. Use when the user wants to "build a loop", "design a loop", "run X continuously/24-7", "make an agent keep working on Y",…