A directory names its project; a compaction is told what to keep; list rows say what a record is #161
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: bvandeusen/FabledScribe#161
Reference in New Issue
Block a user
Three threads, all reaching a session only from
main.#4085 — a directory says which project it belongs to, git repo or not
All six hooks scoped their requests one way:
git remote get-url origin. That key does not exist outside a git repo, so a session in a plain directory was unscoped in every hook at once — no project context, no prior-art scoping, no project rules — and silently, because a missing remote is indistinguishable from a remote nobody bound.A
.scribefile is the second key, read by the sharedscribe_scope_queryso a directory scopes the same way everywhere:instanceis why the file is not just a number: an id is a different project on every Scribe, so a marker that travels — a copied directory, a shared machine, a repo someone else clones — would otherwise scope the session to the wrong project without a word. Compared host-only; a mismatch drops the id, because no project beats the wrong project. A bare integer is accepted too, since that is what a person writes by hand. The marker beats a git remote, which is also how a directory overrides its binding.Two bugs found on the way:
project_idthat did not resolve rendered no message at all — the branch hung offif project_id:as anelif. Harmless while that arg was only used for curl testing; the first thing a marker would hit.$( ). The assignment died with the subshell, leaving the caller to read an unset variable underset -u— which aborts the hook and costs the session its whole SessionStart context, to fetch a warning about a file.#3680 — a PreCompact hook tells the summarizer what must survive
Read out of the installed Claude Code build (2.1.273), not the docs, which describe a different mechanism.
hookSpecificOutput.additionalContextis never read on PreCompact — the schema has no variant for it. But a PreCompact hook's stdout becomes the compaction's custom instructions, merged into the prompt that writes the summary, on every path: manual, auto and partial.So the hook does not interrupt the compaction — it steers the summary, which is what the next turn reads. Blocking is the thing not to do: it skips compaction, tells the model nothing, and leaves the session running on uncompacted with no summary at all.
scribe_precompact_preserve.shnames what the summary is the only copy of: Scribe record ids with titles, the in-progress task and its milestone, work done but not yet recorded, governing rules, and unfinished operator asks. No network, no config, no block on any path.#4061 — list tools return rows that say what a record is, not what it says
list_taskswas overflowing the client at 93k characters, because listing rows carried full bodies. Rows now say what a record is and where it sits.Verification
CI run 6923 on
dd08f98: Python lint, TypeScript typecheck, Python tests, Plugin hooks and integration all green.Not verified live, and cannot be until this merges: the plugin ships from
main, so the marker and the compaction hook both need a/pluginupdate before a real session exercises them.🤖 Generated with Claude Code
https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy
The hook's header quotes `{"decision":"block"}` while explaining why this hook must never emit one, and the instructions it prints use "decision" in a sentence. Grepping the whole file caught both and failed the guard on the file doing its job. Strips comments and the heredoc first, and guards the stripper: `_code()` returning nothing would make all three static checks pass against an empty string, which is the circularity rule 167 is about. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01821k5B3Ysecp9fNYs92KuyAll six hooks scoped their requests one way: `git remote get-url origin`, resolved server-side through the repo bindings. That key does not exist outside a git repo, so a session in a plain directory was unscoped in every hook at once — no project context, no prior-art scoping, no project rules — and silently, because a missing remote is indistinguishable from a remote nobody bound. A `.scribe` file is the second key, read by the shared scribe_scope_query so a directory scopes the same way everywhere: {"instance": "https://scribe.example.com", "project_id": 2, "project": "…"} `instance` is why the file is not just a number: an id is a different project on every Scribe, so a marker that travels — a copied directory, a shared machine, a repo someone else clones — would otherwise scope the session to the wrong project without a word. Compared host-only, and a mismatch drops the id: no project beats the wrong project. A bare integer is accepted too, since it is what a person writes by hand. The marker beats a git remote — someone put the file there on purpose — which is also how a directory overrides its binding. Two things it found on the way: * An explicit project_id that did not resolve rendered NO message at all — the branch hung off `if project_id` as an `elif`, so a caller holding a pointer it believed in got a context that silently omitted the project it had asked for. Now reported. * The refusal reason was a global set inside a function every caller reads through `$( )`. The assignment died with the subshell, leaving the caller to read an unset variable under `set -u` — which aborts the hook and costs the whole session's SessionStart context, to fetch a warning about a file. It comes back through stdout with the id instead, and a test pins it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy