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/masterleopold/book-capture/ocrgit clone --depth 1 https://github.com/masterleopold/book-captureWhat 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 | $0.00014 | $0.00488 |
| Opus 5 | $0.00007 | $0.00244 |
| Sonnet 5 | $0.00003 | $0.00098 |
| Haiku 4.5 | $0.00001 | $0.00049 |
Grade A, and why
ocr scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
OCR Text Extraction
Run OCR on existing page captures. Requires page_* image files (WebP, or PNG for books captured before the WebP switch) to already exist in the captures directory.
Setup
PLUGIN_SCRIPTS=${CLAUDE_PLUGIN_ROOT}/scripts- Resolve
CAPTURES_BASEfrom.claude/book-capture.local.mdor defaultBooks/files/book-captures - Parse
$ARGUMENTSfor BookID OUTPUT_DIR=<CAPTURES_BASE>/<BookID>
Verify captures exist:
Glob: <OUTPUT_DIR>/page_*.{webp,png}
If no captures found, tell the user to run /book-capture:capture first.
Step 1: Vision OCR
bash "${CLAUDE_PLUGIN_ROOT}/scripts/setup.sh" --check-only || bash "${CLAUDE_PLUGIN_ROOT}/scripts/setup.sh"
node "${CLAUDE_PLUGIN_ROOT}/scripts/extract-text.mjs" "<OUTPUT_DIR>" --concurrency 5 --claude-threshold 0.6
Report stats from raw_text.json.
Step 2: Agent Re-reading (if needed)
Check stats in raw_text.json. If vision-low or failed pages exist, ask user if they want agent re-reading.
Generate a unique SESSION_ID (first 8 chars of random hex) to avoid temp file collisions.
Dispatch 4-8 parallel agents (subagent_type: book-capture:ocr-reader), each handling ~20-30 pages. Write to /tmp/ocr_batch_<SESSION_ID>_NN.json.
After agents complete, merge batch results from /tmp/ocr_batch_<SESSION_ID>_*.json back into raw_text.json. Clean up temp files.
Step 3: Report
Report final OCR stats:
- Total pages
- Vision (good) count
- Vision (low) count
- Agent-read count
- Failed count
- Average confidence
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 · 53 lines · 14 tokens per session scan A 3ff415120fac
ocr is a command published in the GitHub repository masterleopold/book-capture (10 stars, last pushed 1mo ago), licensed MIT. It adds 14 tokens to every session and 488 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 2026-08-31.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
constitution
Create or update the project constitution from interactive or provided principle inputs.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.