api passthrough: endpointSuggestions misses the contents API — raw /contents/ calls aren't steered to madt_files #43
Labels
No labels
breaking
bug
documentation
enhancement
epic
good first issue
help wanted
refactoring
resolution/duplicate
resolution/invalid
resolution/wontfix
security
severity/critical
severity/high
severity/low
severity/medium
status/abandoned
status/blocked
status/needs-decision
status/needs-info
status/needs-verification
testing
upstream
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
sixfold-space/madtea#43
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Live evidence (2026-07-13, during the release-recovery work): I made four raw
madt_api_callrequests that all have first-classmadt_filescoverage, and the passthrough steered none of them:GET /repos/{owner}/{repo}/contents/{dir}×2 (directory list →madt_files action=list)GET /repos/{owner}/{repo}/contents/{file}(file read →madt_files action=get)POST /repos/{owner}/{repo}/contents/{file}(file create/commit →madt_files action=create)endpointSuggestions(internal/service/api/suggestions.go) exists precisely so raw-API fallthrough points agents at the typed surface — but it has no mapping for the contents endpoints, so the calls pass through silently and the agent never learnsmadt_filesexists.Ask: add contents-API mappings —
GET|POST|PUT|DELETE /repos/{owner}/{repo}/contents/*(file and directory forms, incl. the bare/contents/root listing) → the correspondingmadt_filesaction, on both surfaces (madtea apiCLI andmadt_api_call).Acceptance:
madt_filesaction.Also absorbed here: the repo previously carried a session-local rule blocking raw
madtea apiuntil an issue was filed for the missing command (removed in #45 — it never reached real users, since repo-local session config only loads inside a madtea clone). The suggestion surface this issue asks for is the shipped mechanism that intent actually needs: every unmapped passthrough call should point at the typed surface or invite an issue for the gap.