Outcome
Run the deterministic copy QA on sequence email steps: either a saved version (sequenceId + versionId) or raw steps [{subject, body, threadWithPrevious?, manualTask?, templateVariables?}], not both; manual LinkedIn/call task steps get no findings, and WhatsApp template steps only an unknown_merge_field error for a variable that is not a merge field. Returns findings per step (code, severity error|warning, message, location) and pass=false when any error exists. Errors: unknown or malformed merge fields, empty subject or body, placeholder text ([Name], TODO, lorem), link shorteners. Warnings: optional merge fields with no fallback, subject over 78 characters, body under 25 or over 200 words, spam-trigger phrases, ALL CAPS words, exclamation marks, more than 2 links, unsubscribe wording in the body (the compliance footer owns opt-out), an unpersonalized first step and repeated subjects. Publishing is refused while errors remain. Read-only; it flags problems and never rewrites copy.
Before you start
- Verify the intended environment, signed-in credential, authorized organization and the action permission workspace:read.
- Discover and select an authorized canonical Engage workspace. Pass resourceType=workspace and resourceId separately from action input.
Discover the target
topolo whoami --json
topolo actions get app_topolo_engage.sequences.copy_check --json
topolo workspaces list topolo-engage --json
Set ENGAGE_WORKSPACE_ID to the identifier of the selected authorized workspace.
Input contract
| Field | Required | Declared constraints |
|---|---|---|
sequenceId | No | {"type":"string","minLength":1,"maxLength":128} |
versionId | No | {"type":"string","minLength":1,"maxLength":128} |
steps | No | {"minItems":1,"maxItems":10,"type":"array","items":{"type":"object","properties":{"subject":{"type":"string","maxLength":200},"body":{"type":"string","maxLength":10000},"threadWithPrevious":{"type":"boolean"},"manualTask":{"type":"boolean"},"templateVariables":{"maxItems":10,"type":"array","items":{"type":"string","maxLength":64}}},"required":["subject","body"],"additionalProperties":false}} |
Validate, plan and execute
Build payload.json from the current schema and discovered identifiers. The action description above defines its behavior, limits, side effects and declared errors.
: "${ENGAGE_WORKSPACE_ID:?Set ENGAGE_WORKSPACE_ID to your selected workspace identifier}"
topolo actions validate app_topolo_engage.sequences.copy_check --data @payload.json --json
topolo actions plan app_topolo_engage.sequences.copy_check --resource-type workspace --resource-id "$ENGAGE_WORKSPACE_ID" --data @payload.json --json
Execute the unchanged validated read.
: "${ENGAGE_WORKSPACE_ID:?Set ENGAGE_WORKSPACE_ID to your selected workspace identifier}"
topolo actions call app_topolo_engage.sequences.copy_check --resource-type workspace --resource-id "$ENGAGE_WORKSPACE_ID" --data @payload.json --json
For MCP, use topolo_get_action, topolo_validate_action, topolo_plan_action and topolo_call_action. Pass resource context separately where required; confirm:true is required for mutations.
Captured successful requests and responses
Captured HTTP 200 · sequences.copy_check and publish enforcement > checks raw steps for a reader and refuses callers without workspace:read
Action input
{
"steps": [
{
"subject": "Hi",
"body": "TODO"
}
]
}
Actual local response
{
"pass": false,
"errorCount": 1,
"warningCount": 2,
"steps": [
{
"position": 1,
"findings": [
{
"code": "body_too_short",
"severity": "warning",
"message": "1 words; cold outreach under 25 words often lacks context.",
"location": {
"field": "body"
}
},
{
"code": "placeholder_text",
"severity": "error",
"message": "Placeholder text \"TODO\" is still in the copy.",
"location": {
"field": "body",
"match": "TODO"
}
},
{
"code": "first_step_not_personalized",
"severity": "warning",
"message": "The first email uses none of {{first_name}}, {{last_name}}, {{company}}, {{title}}.",
"location": {
"field": "body"
}
}
]
}
]
}
Source test: sequences.copy_check and publish enforcement > checks raw steps for a reader and refuses callers without workspace:read. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 200 · sequences.copy_check and publish enforcement > gives raw manual task steps no findings, as the editor sends them
Action input
{
"steps": [
{
"subject": "",
"body": "",
"manualTask": true
}
]
}
Actual local response
{
"pass": true,
"errorCount": 0,
"warningCount": 0,
"steps": [
{
"position": 1,
"findings": []
}
]
}
Source test: sequences.copy_check and publish enforcement > gives raw manual task steps no findings, as the editor sends them. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 200 · sequences.copy_check and publish enforcement > checks a saved version, scoped to the workspace
Action input
{
"sequenceId": "eng_seq_1",
"versionId": "eng_seqv_1"
}
Actual local response
{
"pass": true,
"errorCount": 0,
"warningCount": 6,
"steps": [
{
"position": 1,
"findings": [
{
"code": "merge_field_without_fallback",
"severity": "warning",
"message": "{{first_name}} is optional contact data and has no fallback: contacts without it stop instead of receiving this step.",
"location": {
"field": "subject",
"match": "{{first_name}}"
}
},
{
"code": "body_too_short",
"severity": "warning",
"message": "8 words; cold outreach under 25 words often lacks context.",
"location": {
"field": "body"
}
},
{
"code": "merge_field_without_fallback",
"severity": "warning",
"message": "{{first_name}} is optional contact data and has no fallback: contacts without it stop instead of receiving this step.",
"location": {
"field": "body",
"match": "{{first_name}}"
}
},
{
"code": "merge_field_without_fallback",
"severity": "warning",
"message": "{{company}} is optional contact data and has no fallback: contacts without it stop instead of receiving this step.",
"location": {
"field": "body",
"match": "{{company}}"
}
}
]
},
{
"position": 2,
"findings": [
{
"code": "body_too_short",
"severity": "warning",
"message": "3 words; cold outreach under 25 words often lacks context.",
"location": {
"field": "body"
}
},
{
"code": "merge_field_without_fallback",
"severity": "warning",
"message": "{{first_name}} is optional contact data and has no fallback: contacts without it stop instead of receiving this step.",
"location": {
"field": "body",
"match": "{{first_name}}"
}
}
]
}
]
}
Source test: sequences.copy_check and publish enforcement > checks a saved version, scoped to the workspace. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 200 · WhatsApp template steps in sequences > stores the template with its variables, copies it into a new version and gets no email copy findings
Action input
{
"sequenceId": "eng_seq_1",
"versionId": "eng_seqv_1"
}
Actual local response
{
"pass": true,
"errorCount": 0,
"warningCount": 3,
"steps": [
{
"position": 1,
"findings": [
{
"code": "merge_field_without_fallback",
"severity": "warning",
"message": "{{first_name}} is optional contact data and has no fallback: contacts without it stop instead of receiving this step.",
"location": {
"field": "subject",
"match": "{{first_name}}"
}
},
{
"code": "body_too_short",
"severity": "warning",
"message": "8 words; cold outreach under 25 words often lacks context.",
"location": {
"field": "body"
}
},
{
"code": "merge_field_without_fallback",
"severity": "warning",
"message": "{{first_name}} is optional contact data and has no fallback: contacts without it stop instead of receiving this step.",
"location": {
"field": "body",
"match": "{{first_name}}"
}
}
]
},
{
"position": 2,
"findings": []
}
]
}
Source test: WhatsApp template steps in sequences > stores the template with its variables, copies it into a new version and gets no email copy findings. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 200 · WhatsApp template steps in sequences > copy check flags template variables that are not merge fields
Action input
{
"steps": [
{
"subject": "",
"body": "",
"templateVariables": [
"first_name",
"phone"
]
}
]
}
Actual local response
{
"pass": false,
"errorCount": 1,
"warningCount": 0,
"steps": [
{
"position": 1,
"findings": [
{
"code": "unknown_merge_field",
"severity": "error",
"message": "Template parameter {{2}} maps to \"phone\", which is not a supported merge field (first_name, last_name, company, title, sender_name).",
"location": {
"field": "body",
"match": "phone"
}
}
]
}
]
}
Source test: WhatsApp template steps in sequences > copy check flags template variables that are not merge fields. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Effects and response meaning
- Run the deterministic copy QA on sequence email steps: either a saved version (sequenceId + versionId) or raw steps [{subject, body, threadWithPrevious?, manualTask?, templateVariables?}], not both; manual LinkedIn/call task steps get no findings, and WhatsApp template steps only an unknown_merge_field error for a variable that is not a merge field. Returns findings per step (code, severity error|warning, message, location) and pass=false when any error exists. Errors: unknown or malformed merge fields, empty subject or body, placeholder text ([Name], TODO, lorem), link shorteners. Warnings: optional merge fields with no fallback, subject over 78 characters, body under 25 or over 200 words, spam-trigger phrases, ALL CAPS words, exclamation marks, more than 2 links, unsubscribe wording in the body (the compliance footer owns opt-out), an unpersonalized first step and repeated subjects. Publishing is refused while errors remain. Read-only; it flags problems and never rewrites copy.
Verify the outcome
- Check the response against the declared output schema and the requested operation. For mutations, use the action-specific read described in the action description before any dependent operation.
Recover
For an uncertain result, inspect current retained state with a read action before retrying. Do not replay a mutation unless its contract explicitly guarantees idempotency.
Refused requests
Captured HTTP 403 · sequences.copy_check and publish enforcement > checks raw steps for a reader and refuses callers without workspace:read
Action input
{
"steps": [
{
"subject": "Hi",
"body": "x"
}
]
}
Actual local response
{
"error": {
"code": "permission_denied",
"message": "The credential lacks workspace:read.",
"details": {
"requiredPermission": "workspace:read"
}
}
}
Source test: sequences.copy_check and publish enforcement > checks raw steps for a reader and refuses callers without workspace:read. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 404 · sequences.copy_check and publish enforcement > checks a saved version, scoped to the workspace
Action input
{
"sequenceId": "eng_seq_1",
"versionId": "eng_seqv_1"
}
Actual local response
{
"error": {
"code": "not_found",
"message": "Sequence version was not found."
}
}
Source test: sequences.copy_check and publish enforcement > checks a saved version, scoped to the workspace. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 422 · sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps
Action input
{}
Actual local response
{
"error": {
"code": "invalid_action_input",
"message": "Pass either sequenceId and versionId, or steps — not both."
}
}
Source test: sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 422 · sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps
Action input
{
"sequenceId": "s"
}
Actual local response
{
"error": {
"code": "invalid_action_input",
"message": "Pass either sequenceId and versionId, or steps — not both."
}
}
Source test: sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 422 · sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps
Action input
{
"sequenceId": "s",
"versionId": "v",
"steps": [
{
"subject": "a",
"body": "b"
}
]
}
Actual local response
{
"error": {
"code": "invalid_action_input",
"message": "Pass either sequenceId and versionId, or steps — not both."
}
}
Source test: sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Captured HTTP 422 · sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps
Action input
{
"versionId": "v",
"steps": [
{
"subject": "a",
"body": "b"
}
]
}
Actual local response
{
"error": {
"code": "invalid_action_input",
"message": "Pass either sequenceId and versionId, or steps — not both."
}
}
Source test: sequences.copy_check and publish enforcement > requires exactly one of a version or raw steps. Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
Error recovery
- HTTP 403
permission_denied: Stop and inspect the current credential and selected resource. It needs workspace:read; do not broaden access or switch organization automatically. - HTTP 404
not_found: Rediscover the identifier using the related list or get action in the same authorized resource. Do not guess identifiers or change organization context. - HTTP 422
invalid_action_input: Correct the request against the current action input schema, then validate it again before planning.
Related actions
- List sequences
- Create sequence
- Get sequence
- Rename sequence
- Archive sequence
- Create sequence draft version
- Get sequence version
- Set draft sequence steps
- Publish sequence version
- Draft sequence steps with AI
Evidence and limits
-
/api/sequences/copy-check -
Actual route handlers with SQLite repositories; synthetic identity/workspace grants and provider ports.
-
Widget/bootstrap use actual app handler but mocked worker runtime and app identity.
-
No gateway execution, live permissions, external delivery or deployed proof.
These are local route captures with synthetic identity and simulated dependencies. They do not prove live authorization, gateway execution, production state, or external delivery.