site: reposition the landing page around the open session format - #180
Merged
Conversation
New hero tagline, a two-front problem statement (locked-in session logs, then provenance lost at merge), a supported-harness matrix, persona-led scenario cards (cross-harness resume, one query surface), a quick start that leads with cache sync / query / share / resume, and a workspace table covering all published crates.
Hero leads with the four-word claim and unpacks portability in the next sentence; the problem section closes on the same phrase; the site-wide meta description follows.
Frame parsing in and projecting out around the format's stability: harness formats are undocumented and change without notice, so they are treated as boundaries around a versioned schema.
One short paragraph sets the model (versioned center, fragile edges); parse in, project out, and resume build on it as labeled bullets.
Page order becomes problem, scenarios, core objects, then the matrix and boundary model, then the git comparison.
The versioned-open rhetoric and the three moves return to just after the problem statement under a Parse in, project out heading; the harness matrix stays below the fold with a one-line caption.
The git and GitHub one-graph note moves down to the supported-harness section, next to the git comparison.
Each card opens on a recognizable moment and closes on the move that fixes it. Also trim the hero's top padding (2xl to lg desktop, xl to md narrow) so the title sits higher.
Replace the harness-switching card with the ship-the-session-with- the-PR case: reviewers get the asked/tried/verified record, so merge- stalling questions are answered before they're asked.
Cross-device, cross-person, cross-harness handoff replaces the multi-actor PR card, whose ground is already covered by the problem statement and the git comparison table.
Drop the all-checkmark read/write/resume columns; end-to-end support becomes one sentence. The table now answers the two questions a visitor brings: can I trust what it says changed, and what it cost. Cells grade answer quality per harness; Copilot is tagged preview.
Readers scan for their harness's name and the exceptions; parse and project quality become one strong claim each, with per-harness gaps delegated to the format notes.
|
🔍 Preview deployed: https://93fc51b2.toolpath.pages.dev |
Flex-wrapped labels on the card surface with a copper hover, caveats set small in secondary; replaces the bulleted grid styling.
Replace the fourteen-row crate table with three pointers: the spec for any-language implementers, the crates for library consumers, and the Claude Code plugin for in-session use. The crates page keeps the full listing.
Fold format fragility into its closing sentence; the git-blame and fixes-both paragraphs go, since the boundary model and comparison table cover that ground.
The definitional claim returns between the problem and the scenarios, paired with the N-plus-M argument: tools build against the format once, and a new agent needs one parser to join every tool.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The landing page pitched Toolpath provenance-first: "what happens
between commits", git blame, merge-time archaeology. Most use now is
portable sessions: getting them out of a harness, moving them across
machines, people, and harnesses, and building tooling against one
schema instead of one parser per agent. The copy didn't say that until
several screens in, if at all.
This rewrites the page to lead with the portable-session story: the
open, versioned format as the stable center, undocumented harness
formats as the fragile edges, and parse in / project out / resume as
the three moves across that boundary. Provenance stays, but as a
supporting act (the comparison table and the review-time card) rather
than the frame.
Along the way: the hero, problem statement, and scenario cards are
rewritten around that argument; the harness support presentation drops
all-checkmark columns in favor of a chip list linking each harness to
its crate docs; and the closing fourteen-row crate table becomes three
next-step pointers (spec, crates, plugin), with the full listing left
to /crates/. Hero top padding is trimmed and the site meta description
updated to match.
pnpm run buildproduces the expected 11 pages.