Issues

`hook-guard read` skips any path containing a directory with the same name as the last segment of `GRAPHIFY_OUT`

#3959

· chang258 opened

on Oct 1, 2026

semantically_similar_to is scoped to one chunk, so similarity across chunks is never emitted, and plain string metrics miss it in a reconciliation pass

#3948

· jnrod03-rgb opened

on Sep 30, 2026

extraction-spec has no rule for referencing nodes outside the chunk: agents copy id fragments or mint stub nodes, and replace-on-re-extract deletes real content

#3947

· jnrod03-rgb opened

on Sep 30, 2026

--update: break down a mass change by top-level directory before dispatching extraction

#3945

· jnrod03-rgb opened

on Sep 30, 2026

Code files that yield no symbols (data-only .js) enter the graph as isolated nodes, silently

#3946

· jnrod03-rgb opened

on Sep 30, 2026

update.md prints the whole detect_incremental result to stdout, flooding the agent's context

#3944

· jnrod03-rgb opened

on Sep 30, 2026

Windows + uv: Step 1 pins an ephemeral uv cache interpreter, and the subcommand guard still parses graphify.exe as a shebang

#3943

· jnrod03-rgb opened

on Sep 30, 2026

AST extraction: Svelte 5 script blocks and TypeScript export type * produce partial graphs

#3942

· noxlovette opened

on Sep 30, 2026

SKILL.md never tells the user that graphify-out/ was written into their working tree

#3941

· bonatoriccardo opened

on Sep 30, 2026

Svelte: every .svelte file reports a syntax error at line 1; functions inside <script> are never extracted (extract_svelte feeds the whole file to the JS grammar)

#3928

· chrisltq97 opened

on Sep 29, 2026

[Bug]: inconsistent agent instructions

bug

#3925

· SvenMeyer opened

on Sep 29, 2026

MCP server: silent wrong matches on ambiguous/near-miss names, noisy query_graph, misleading list_prs error

#3917

· Alexandre-Aktisea opened

on Sep 29, 2026

Navigation Menu

Uh oh!

All issues

`hook-guard read` skips any path containing a directory with the same name as the last segment of `GRAPHIFY_OUT`

semantically_similar_to is scoped to one chunk, so similarity across chunks is never emitted, and plain string metrics miss it in a reconciliation pass

extraction-spec has no rule for referencing nodes outside the chunk: agents copy id fragments or mint stub nodes, and replace-on-re-extract deletes real content

--update: break down a mass change by top-level directory before dispatching extraction

Code files that yield no symbols (data-only .js) enter the graph as isolated nodes, silently

update.md prints the whole detect_incremental result to stdout, flooding the agent's context

Windows + uv: Step 1 pins an ephemeral uv cache interpreter, and the subcommand guard still parses graphify.exe as a shebang

AST extraction: Svelte 5 script blocks and TypeScript export type * produce partial graphs

SKILL.md never tells the user that graphify-out/ was written into their working tree

Svelte: every .svelte file reports a syntax error at line 1; functions inside <script> are never extracted (extract_svelte feeds the whole file to the JS grammar)

[Bug]: inconsistent agent instructions

MCP server: silent wrong matches on ambiguous/near-miss names, noisy query_graph, misleading list_prs error

Uh oh!

Search results