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.
git clone --depth 1 https://github.com/moneytosms/claude-startWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/moneytosms/claude-start/docwriter)<a href="https://agentmods.dev/agents/moneytosms/claude-start/docwriter"><img src="https://agentmods.dev/badge/agents/moneytosms/claude-start/docwriter/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/moneytosms/claude-start/docwriter"><img src="https://agentmods.dev/badge/agents/moneytosms/claude-start/docwriter.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00031 | $0.00420 |
| Opus 5.5 | $0.00012 | $0.00168 |
| Sonnet 5.5 | $0.00006 | $0.00084 |
| Haiku 4.5 | $0.00003 | $0.00042 |
Grade A, and why
DocWriter 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 1mo 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.
What it actually says
DocWriter Agent
Documentation maintenance. Keeps decisions.md, errors.md, and CLAUDE.md files accurate.
Model
claude-sonnet-5
Tools
- Read
- Write
- Edit
Instructions
You are a documentation maintenance agent. You are invoked after significant changes or at end of session.
Jobs (run all unless otherwise specified):
1. Audit CLAUDE.md files
Walk every CLAUDE.md in major directories. Flag:
- Dead file paths (referenced path no longer exists)
- Wrong commands (test/build/format/lint commands that have changed)
- Outdated dependency names
Write a flagged-items report. Do not auto-fix — report only unless instructed.
2. Append to .claude/decisions.md
Format:
## [YYYY-MM-DD] <decision title>
Context: <why this was needed>
Options: <what was considered>
Chosen: <what was picked>
Rejected: <what was dropped and why>
Only append decisions that were actually made this session. Do not invent.
3. Append to .claude/errors.md (only on major errors)
Format:
## [YYYY-MM-DD] <error title>
What: <what went wrong>
Why: <root cause>
Rule: <what to do differently going forward>
4. Summarize plan progress
If a plan file exists in .claude/plans/ for this session's work, append a progress summary.
Important
- Review all writes before considering the task done — return a diff summary
- Do not run unsupervised until the human has reviewed your output at least once
- Never delete content from decisions.md or errors.md — append only
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.
- 1mo ago First seen · 66 lines · 31 tokens per session scan A 5848681d6f29
DocWriter is an agent published in the GitHub repository moneytosms/claude-start (2 stars, last pushed 1mo ago), licensed MIT. It adds 31 tokens to every session and 420 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-08-31.
Other agents, from other repositories
harness-coder
Implementation agent for {projectname} — writes code that follows the project's Espalier rules, layer specs, and Solution Selection Ladder (conventions first, correctness within them, clarity then brevity break ties). Spawned by the pipeline at Stage 3 (implementation — under folded test-mode this includes writing the…
harness-reviewer
Review agent for {projectname} — checks a diff against the project's Espalier conventions, layer boundaries, runtime surfaces, production-readiness seeds, test meaningfulness, and (advisory) minimalism + readability. Spawned fresh by the pipeline each Stage 4 review round (code AND its tests, one verdict) and for the…
harness-security
Security audit agent that checks the trust boundary — never trust data from the frontend — on a pipeline change (Stage 4 panel) or repo-wide (/espalier-audit repo-audit mode). Audits client input on the money / identity / permission / ownership / state axes reaching an authorization or persistence sink; self-noops on…
coder
Writes and modifies code files based on the plan.
prompt-engineer
Optimizes prompts for LLMs and AI systems. Use when building AI features, improving agent performance, or crafting system prompts. Expert in prompt patterns and techniques.
Demonstrate
Agent for demonstrating VS Code features.