Hosted onhyper.mediavia theHypermedia Protocol
Develop Seed
Mentions of Clerk
Respond to the mention, performing the action requested.
One-shot Reed ch5 retry
Process ~/memory/imports/reed-sections/ch5.md completely and copy-edit it in place: faithful historical transcription of David P. Reed's 1978 dissertation; correct only clear OCR corruption, malformed words, stray scan glyphs, misread letters/digits, broken citation brackets, obvious spacing artifacts, and clearly broken Markdown; preserve wording, meaning, terminology, spelling conventions, citation style, every heading/link/image/code passage/blockquote, and every exact <!-- id:... --> identity comment; section levels # chapter, ## 5.X, ### 5.X.Y; never remove substantive text or guess uncertain readings. IMPORTANT: direct reads truncate at 8KB, so immediately delegate a SCRIPT child. The script must use ctx.call('read',{address:'~/memory/imports/reed-sections/ch5.md'}) to get full content internally, split at paragraph boundaries into manageable chunks, send all chunks in parallel via ctx.delegate for conservative editing with structured exact text outputs, recombine in order, validate IDs and size, ctx.call('write',...) the complete revised content to the exact same path, then ctx.call('read',...) and verify equality/completeness. After success write ~/memory/imports/reed-sections/ch5.copyedit.done.md with counts, correction categories, and validation outcome. Do not merely report instructions; ensure the file is actually saved.
One-shot copyedit Reed ch4
Copy-edit the complete file ~/memory/imports/reed-sections/ch4.md in place for the current user request. This is a faithful historical transcription of David P. Reed's 1978 dissertation, not a rewrite. Correct only clear OCR corruption (malformed words, stray scan glyphs, misread letters/digits, broken citation brackets, obvious spacing artifacts). Preserve Reed's wording, meaning, terminology, spelling conventions, and citation style; do not modernize or summarize. Preserve every Markdown heading, link, image, code passage, blockquote, and every exact <!-- id:... --> identity comment. Improve Markdown only where clearly broken; heading levels are # chapter, ## 4.X, ### 4.X.Y. Do not remove substantive text or invent uncertain text; leave uncertain readings unchanged. Process the ENTIRE ~43 KB file using bounded reads, not only the initial truncated output. Save the full revision to the same exact path. Re-read and validate that the saved file is complete and not truncated, and that structural/identity elements are preserved. After successful completion, write a concise completion marker to ~/memory/imports/reed-sections/ch4.copyedit.done.md with before/after byte and line counts if available, categories of corrections, and validation results. Do not paste the chapter.
One-shot copyedit Reed ch5
Copy-edit the complete file ~/memory/imports/reed-sections/ch5.md in place for the current user request. Read the entire file, correct only clear OCR corruption and clearly broken Markdown, preserve Reed's exact wording/meaning/terminology/spelling/citation style, preserve every heading/link/image/code passage/blockquote and every exact <!-- id:... --> identity comment, ensure heading levels # chapter, ## 5.X, ### 5.X.Y, do not invent or remove text, and process the whole file. Use a visible 3-step plan. Because the file is large, delegate chunks in parallel if useful, then reassemble and validate identities/links/completeness before saving. Re-read the saved file enough to confirm it is not truncated. After successful completion write a short completion marker to ~/memory/imports/reed-sections/ch5.copyedit.done.md containing saved byte/character counts, correction categories, and validation outcome.
One-shot Reed ch2 retry
Complete the user's copy-edit request for ~/memory/imports/reed-sections/ch2.md and actually save it in place. Use ONE script child that internally reads the full file via ctx.call('read'), splits it at paragraph boundaries into ~6000-character chunks, edits all chunks concurrently with ctx.parallel + ctx.delegate, validates that every exact <!-- id:... --> sequence is unchanged, recombines in order, writes the complete output to the same memory path, re-reads it internally, and verifies equality/completeness. Conservative faithful historical transcription: only clear OCR corruption, citation/spacing artifacts, and broken Markdown; headings # chapter, ## X.Y, ### X.Y.Z; no modernization or guessing. The script must return saved bytes, ID count, and verification. Do not merely describe limitations; execute and save.
One-shot copyedit Reed ch2
Process ~/memory/imports/reed-sections/ch2.md completely and copy-edit it in place according to this specification: faithful historical transcription of David P. Reed's 1978 dissertation; correct only clear OCR corruption, malformed words, stray scan glyphs, misread letters/digits, broken citation brackets, and obvious spacing artifacts; preserve wording, meaning, terminology, spelling conventions, citation style, every Markdown heading/link/image/code passage/blockquote, and every exact <!-- id:... -->; fix clearly broken Markdown; section levels # for chapter, ## for X.Y, ### for X.Y.Z; do not remove substantive text or guess uncertain readings; reread after saving to verify completeness. IMPORTANT: because direct reads truncate at 8KB and code tools may be disabled, delegate a SCRIPT child. Have the script use ctx.call('read', {address:'~/memory/imports/reed-sections/ch2.md'}) to obtain the full content internally, split it only at paragraph boundaries into manageable overlapping/nonoverlapping chunks, send all chunks in parallel to model children via ctx.delegate for conservative OCR copy-editing (require exact output text and preservation of IDs), recombine in original order, ctx.call('write', ...) the complete content to the same path, then ctx.call('read',...) and verify byte/ID/heading completeness. Ensure the file is actually saved. Return a concise completion report.
PDF Importer
Prepare PDF-to-Seed import jobs with explicit destination, conversion route, fidelity QA, IPFS asset handling, and citation verification steps.
Do you like what you are reading? Subscribe to receive updates.
Unsubscribe anytime