| name | publish-blog |
| description | Render surviving research findings into a publishable blog post through the typed findings-to-artifact contract. A first-class harness output (not a pack). Use this skill when the user wants to turn research into a blog post, publish findings, draft an article from the corpus, or write up what the research found. Triggers on 'publish blog', 'blog post from research', 'write up these findings', 'draft an article', 'turn this into a post'. |
| version | 0.4.3 |
| argument-hint | [--topic <id>] [--genre <genre>] [--out <path.md>] |
| allowed-tools | Read, Bash, Edit, Glob, Grep |
publish-blog — research → blog post
Blog is the first-class published harness output (SPEC §6d): it ships in the
core, always on, not as a pack (book and other channels arrive via optional channel
packs). It renders a published post from the surviving
findings of a session through the typed findings→artifact contract
(schemas/artifact.schema.json) — the same contract the book output uses, so the
citation-integrity and citation-leak gates run uniformly across both.
Pipeline
-
Resolve the finding set. Default to the topic's findings dir
(reports/<topic>/findings/). Only surviving findings (verdict ≠
falsified) are eligible — quarantined and falsified findings never publish.
-
Synthesize the artifact. Run the report-synthesizer's substrate:
scripts/synthesize-artifact.sh reports/<topic>/findings <genre> reports/<topic>/artifact.json
The artifact is genre/channel-neutral and validates against
schemas/artifact.schema.json. The newsworthiness field carries the
engine's delta signal — the hook a publication needs.
-
Render the post. Every output lives under reports/<topic>/, never a
top-level blog/ directory — the trailing .blog.md suffix (not a leading
blog/ prefix) is what identifies the channel:
scripts/render-artifact.sh reports/<topic>/artifact.json blog reports/<topic>/<topic>.blog.md
Requesting a specific genre alongside the blog channel renders to
reports/<topic>/<topic>.<genre>.blog.md instead (keeps it collision-free
against the report channel's own reports/<topic>/<topic>.<genre>.md
files).
-
Gate the output. The post must read as if written from public primary
sources alone. The bundled check-citation-leak.sh hook fires on write; before
reporting done, confirm the body has zero leaks (no finding/concept ids
urn:mif:concept:/urn:mif:report:, no reports/<slug>/ paths) and that every
claim traces to a ## Sources entry. The post's own urn:mif:blog: frontmatter
id (its MIF Level-1 identity) is not a leak.
Non-negotiables
- Never publish a falsified or quarantined finding. The synthesizer filters
them; do not reintroduce them by hand.
- Never leak internal research references into the prose. Re-author from the
primary source; do not just delete the token.
- The post is a projection of the artifact — to change the content, change the
findings or the synthesis, not the rendered Markdown in place.
- Rebuild the topic README after publishing. The post renders to
reports/<topic>/<topic>.blog.md, inside reports/<topic>/ alongside the
canonical report and any genre-suffixed report-channel files. The
navigation README's Reports table is built from every reports/<topic>/*.md
file (via scripts/build-topic-readme.sh), so a blog write now changes what
that table should show — run bash scripts/build-topic-readme.sh <topic>
(or the readme skill) before reporting this done, the same as
publish-report already does.
blog is a MIF-exempt channel (declared mifExempt: true in
harness.config.json outputs[]): its public prose is orthogonal to MIF, so the
MIF Level-3 source of truth lives in the generic report channel
(reports/<topic>/<slug>.md), not the post. The MIF I/O conformance gate skips
and logs the post; it does not require an L3 projection of it.
