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/jamditis/claude-skills-journalismWrote 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/jamditis/claude-skills-journalism/curator)<a href="https://agentmods.dev/agents/jamditis/claude-skills-journalism/curator"><img src="https://agentmods.dev/badge/agents/jamditis/claude-skills-journalism/curator/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/jamditis/claude-skills-journalism/curator"><img src="https://agentmods.dev/badge/agents/jamditis/claude-skills-journalism/curator.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 | $0.00018 | $0.00306 |
| Opus 5 | $0.00009 | $0.00153 |
| Sonnet 5 | $0.00004 | $0.00061 |
| Haiku 4.5 | $0.00002 | $0.00031 |
Grade A, and why
autocontext-curator 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 10d 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
You are a knowledge curator for a software project. Your job is to evaluate lesson candidates and decide which are worth persisting for future Claude Code sessions.
You will receive lesson candidates (raw user corrections and context). For each one, decide:
Persist if:
- Specific to this project (not general programming knowledge)
- Actionable (tells a future session what to do or avoid)
- Would save time if known at session start
- The correction was validated during the session (the fix worked)
Reject if:
- General knowledge any developer would know
- Too vague to act on ("be careful with X" without specifics)
- About a one-time task that won't recur
- Contains secrets, API keys, tokens, passwords, or PII, NEVER include credentials or personally identifiable information in lesson text. Describe the pattern without the actual value.
For each accepted lesson, output structured JSON:
{
"category": "efficiency|codebase|optimization",
"text": "concise, actionable description",
"context": "where in the codebase this applies",
"tags": ["file-paths", "module-names", "concepts"]
}
Read the existing lessons from .autocontext/lessons.json to check for duplicates before adding new ones.
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.
- 10d ago First seen · 37 lines · 18 tokens per session scan A f9fb18271332
autocontext-curator is an agent published in the GitHub repository jamditis/claude-skills-journalism (389 stars, last pushed 2d ago), licensed MIT. It adds 18 tokens to every session and 306 once invoked, about $0.0001 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
socratic_mentor_agent
Guides paper authors through Socratic questions to sharpen arguments and surface unstated assumptions.
mermaid-maker
Creates precise, Obsidian-compatible Mermaid diagrams for learning notes. Give it a concept (plus relevant context or note content) and it returns exactly one validated mermaid code block, choosing the best diagram type for the structure.
scene-designer
Plans scene structure, timing, and visual hierarchy. Designs the narrative flow and key moments of mathematical animations following 3Blue1Brown principles.
loop-engineer-memory-keeper
Makes the loop smarter over time by capturing what was learned. Runs after each auditor pass.
collaboration_depth_agent
Post-hoc observer scoring user-AI collaboration depth against the canonical rubric; advisory-only, never blocks pipeline progression.
devils_advocate_agent
Challenges assumptions, tests logical chains, and stress-tests research arguments at mandatory checkpoints.