dimagi-internal/ace / skills/solicitation-google-form-qa
solicitation-google-form-qa
VerifiedPass/fail check of a solicitation's public Google Form against the live call.
Solicitation Google Form QA
Structural checks on the form solicitation-google-form made. Binary verdict: pass / fail / incomplete. 8 static checks, run by scripts/qa-run.ts over checks.ts.
It checks the LIVE form and an anonymous fetch, not just the record, because the record says
what the producer meant to make and only the live reads say what a respondent meets. A form is
public, so a defect here is already in front of people. There is no -eval: the form is a
mirror of the solicitation, whose quality solicitation-create-eval grades. See
skills/_qa-template.md for the shared QA contract.
Inputs
| Parameter | Used for | Resolver |
|---|---|---|
record | solicitation-google-form_record.md frontmatter | explicit file id; default: the run's 8-solicitation-management/ folder |
solicitation | live get_solicitation (is_public, questions) | read live from record.solicitation_id |
form | live get_google_form_definition({formId}) | read live from record.form_id |
anonymous_http_status | curl -s -o /dev/null -w '%{http_code}' '<responder_url>', no cookies | measured at QA time |
deny_list | the opp's deny list (inputs/confidentiality.yaml llo_facing_deny_list: plus brief/PDD instructions) | same resolver as the producer; [] when none is declared |
run_products_google_form | run_state phases.solicitation-management.products.solicitation.google_form | run_state when a run is bound; null when none is |
Products
8-solicitation-management/solicitation-google-form-qa_result.yaml, the QA result perlib/qa-types.ts. Outside a run, return it inline.
Checks
| # | id | type | description | auto-fix on fail |
|---|---|---|---|---|
| 1 | record_complete | static | Record carries form_id, edit_url, responder_url, solicitation_id and a non-empty question_map | rewrite the record from the create result |
| 2 | every_question_mirrored | static | question_map covers exactly the live solicitation's question ids plus the three identity items | rebuild the form (new name) from the current call |
| 3 | form_items_carry_refs | static | The live form's items carry unique [ref: …] lines that match question_map | restore the edited item, or rebuild |
| 4 | banner_when_private | static | A private call's form description starts with the demonstration banner | restore the banner |
| 5 | no_deny_listed_party | static | No respondent-facing text on the live form names a deny-listed party | edit the form now, then the call |
| 6 | responder_link_open | static | An anonymous fetch of the responder link returns 200, unless the form is closed | forms_set_publish_settings with anyoneWithLinkCanRespond: true |
| 7 | products_recorded | static | run_state google_form matches the record (skipped when no run is bound) | rewrite the products block |
| 8 | imports_unique | static | No form response or labs response appears twice in imported | delete the duplicate labs response, fix the record |
Process
-
Read the record frontmatter, then
get_solicitation,get_google_form_definitionand the anonymous fetch (§ Inputs). Inside a run, read the products block; outside one, usenull. -
Write the six inputs as one JSON evidence bundle (shape documented at the top of
checks.ts) to a local file. -
Run the checks:
ACE_ROOT="${CLAUDE_PLUGIN_ROOT:-$(python3 -c "import json,os; d=json.load(open(os.path.expanduser('~/.claude/plugins/installed_plugins.json'))); print(d['plugins']['ace@ace'][0]['installPath'])")}" node "$ACE_ROOT/node_modules/tsx/dist/cli.mjs" "$ACE_ROOT/scripts/qa-run.ts" \ --skill solicitation-google-form-qa --artifact <bundle.json> \ --target <opp>/<run-id> \ --capture-path 8-solicitation-management/solicitation-google-form_record.md -
Write the result through the shared writer, never by hand. Put one outcome per check (
{check, type: "static", result: {pass, detail, auto_fix_hint}}) in a localoutcomes.json, then:node "$ACE_ROOT/node_modules/tsx/dist/cli.mjs" "$ACE_ROOT/scripts/qa-result.ts" write \ --skill solicitation-google-form-qa --target <opp>/<run-id> \ --capture-path 8-solicitation-management/solicitation-google-form_record.md \ --outcomes outcomes.json --out <local result.yaml>Upload that file to § Products. A hand-written shape is refused by the ace-gdrive write guard (
skills/_qa-template.md§ QA result YAML contract). -
On
fail, a check 5 or check 6 failure is urgent: the form is public. Fix it before anything else. -
A record that does not exist is
incomplete, notfail: there is no form to check.
MCP Tools Used
- ace-gdrive:
drive_read_file,get_google_form_definition,drive_create_file - connect-labs:
get_solicitation
Mode Behavior
Same in every mode: read, check, write the result. It never changes the form.
Change Log
| Date | Change | Author |
|---|---|---|
| 2026-10-18 | Initial version (optional-track PR5b): 8 checks over the live form, an anonymous fetch and the record | ACE (optional labs E2E) |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |