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 commands/michael-ovo/obsidian-knowledge-agent/polishgit clone --depth 1 https://github.com/Michael-OvO/obsidian-knowledge-agentWrote 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/commands/michael-ovo/obsidian-knowledge-agent/polish)<a href="https://agentmods.dev/commands/michael-ovo/obsidian-knowledge-agent/polish"><img src="https://agentmods.dev/badge/commands/michael-ovo/obsidian-knowledge-agent/polish.svg" alt="Measured on agentmods" 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 | $0.00250 |
| Opus 5 | $0.00009 | $0.00125 |
| Sonnet 5 | $0.00004 | $0.00050 |
| Haiku 4.5 | $0.00002 | $0.00025 |
Grade A, and why
polish scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Run the teaching-quality pass on the target ($ARGUMENTS, or the most recently changed
notes if none is given). Read .agents/style-guide.md first.
- Select the note(s). For a folder, polish the content and concept notes (skip indexes and source notes unless I ask).
- Review against the style guide: does the body read naturally top to bottom? Is the machinery off-stage (metadata in frontmatter, no rubric headings)? For technical notes, is it bullet-first with the artifacts that genuinely teach (runnable code, an equation, a diagram)? Are related links specific and useful?
- Rewrite only the failing parts — never restate a whole note and never change its meaning. Add a missing artifact only where it actually helps.
- Fix and wire wikilinks, then validate them (the bundled
scripts/validate_links.py, or the inline check in the workflow). Show me the diffs before committing.
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.
- 6d ago First seen · 19 lines · 18 tokens per session scan A 1e65581a6ea2
polish is a command published in the GitHub repository Michael-OvO/obsidian-knowledge-agent (206 stars, last pushed 2mo ago), licensed MIT. It adds 18 tokens to every session and 250 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 commands, from other repositories
memory-update
Run the Memory Update protocol. Checks whether decisions, questions, assumptions, risks, tasks, or opportunities emerged in the current task and writes them to the right files, then prints the required end-of-task summary.
install
Guided setup for a new Cowork workspace. Interviews you (about 5 minutes) and generates a full, pre-configured cowork-os workspace (folders, copy, project instructions, and recurring routines).
knowledge-transfer
Interview a person to capture their operating know-how (tacit knowledge) and write it into the workspace as the company brain: processes, rules with source, glossary, decisions and open questions. For onboarding, a key person leaving, or standardizing a process.
linkedin-editor
Run the 8-step LinkedIn editor on any source (an article, a case study, a draft, rough notes) and deliver ready-to-publish, founder-voice LinkedIn posts optimized for reach and leads, not likes.
mission
Start or advance a mission, an ambitious outcome (a dream client, a partnership, breaking into named accounts) that Claude pursues as an outcome, not a task, with route maps, an evidence log, and stop conditions.
process-inbox
Process the inbox. Read everything the user dropped in (notes, links, a deck, a messy folder) and route each fact to the most specific workspace file, separating facts from assumptions and open questions, without inventing anything.