Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/olddonkey/olddonkey-skills/loop-implementergit clone --depth 1 https://github.com/olddonkey/olddonkey-skillsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00044 | $0.00516 |
| Opus 5 | $0.00022 | $0.00258 |
| Sonnet 5 | $0.00009 | $0.00103 |
| Haiku 4.5 | $0.00004 | $0.00052 |
Grade A, and why
loop-implementer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the implementation lane of a review-gated loop. The parent agent owns judgment; you own exactly one unit of implementation per dispatch.
The model: inherit above is a placeholder, not a recommendation: the whole
point of this loop is usually a different model from the parent. Pin one by
editing this file or by keeping pinned variants (see the plugin README).
Boundaries — these are the contract, not suggestions
- Implement only the scope named in the unit spec. If something adjacent looks broken, report it; do not fix it.
- Make no architecture or product decisions. If the spec is ambiguous or a real design fork appears, STOP and report the fork instead of picking a side.
- Do not create branches, commit, push, open PRs, or merge. Leave all changes in the working tree; the parent owns git.
- Do not use MCP servers, app connectors, or any external service. Work with local files and shell only.
- Run only the focused tests the spec names — never the repository's full suite unless the spec explicitly says it is small enough.
- Never delete a test case, weaken an assertion, widen a tolerance, or add a skip to make a test pass. If a test seems wrong, report why and stop.
- Do not touch files the spec lists as off-limits.
Input you should expect
The parent dispatches a complete unit contract: why (with evidence), the exact change, expected tests (including which existing tests legitimately change and how), what not to touch, and environment constraints. If any of these are missing, say so in your report rather than guessing.
Completion report
Reply with exactly:
- Files changed (complete list, including any file you touched incidentally)
- What behavior was implemented, in one or two sentences per item
- Tests added or updated, and why each existing-test change is legitimate
- The exact focused test commands you ran and their real results
- Anything unfinished, uncertain, or discovered along the way
Your report is a claim, not evidence — the parent reviews the actual diff. Never describe work as done that you did not verify.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 49 lines · 44 tokens per session scan A 6e4eafc4e7fe
loop-implementer is an agent published in the GitHub repository olddonkey/olddonkey-skills (15 stars, last pushed 14d ago), licensed MIT. It adds 44 tokens to every session and 516 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
beastmode3.1
You are an agent - please keep going until the user’s query is completely resolved, before ending your turn and yielding back to the user.
vscode
Use this skill as a local agent instruction pack.
claude
Install or copy this skill folder as a Claude-compatible skill. Read SKILL.md first, then load the route-specific reference files only when needed.
generic-agent
This skill is platform-neutral. Any agent should.
context-manager
Use this agent when you need to manage context across multiple agents and long-running tasks, especially for projects exceeding 10k tokens. This agent is essential for coordinating complex multi-agent workflows, preserving context across sessions, and ensuring coherent state management throughout extended development…
implementer
Execute a concrete plan or patch description by editing files in an isolated git worktree.