Check the input, then compare the complete deliverables
Explore the original materials, item-by-item outcomes and full report below. Attempt the input independently before using the reference to find omissions or disagreements.
The article and reference drafts are synthetic. No Typefully connection, draft creation, scheduling or publication was performed.
Original synthetic material. Try the input independently, then compare the reference results.
Original article
# Define the deliverable before automating Author: shancha | Synthetic practice article | 2026-10-17 [F01] This exercise uses 8 synthetic emails covering requirements, waiting, bills, receipts, subscriptions, reference material and insufficient context. They are not from a production inbox. [F02] The rules use 6 business labels plus Review. Each email receives one primary workflow label. [F03] Deliverables are a classification table, 2 reply drafts and a daily brief. Drafts exist on the practice page only; none was saved to Gmail or sent. [F04] Unknown meeting times, prices and delivery dates must remain pending; drafts cannot promise them. [F05] Validate small samples first, inspect real account capabilities separately, then run a small trial before saving a Skill or considering a Routine. [F06] Time savings, accuracy and revenue changes were not measured. Real inbox automation is unverified; do not infer performance numbers. The order matters: define outputs and acceptance, preserve uncertainty, then decide which part is worth repeating. Rewritten content should keep every claim tied to the source; the author decides scheduling and publication.
Editorial brief
X:5 posts; LinkedIn:1 post. Concrete, restrained tone; retain synthetic practice and real workflow unverified. Cite facts. Local files only.
Complete reference results under this tutorial’s rules; nothing was written to a third-party service.Scroll horizontally within the table if needed.
| Draft | Channel | Facts | Reference text |
|---|---|---|---|
| X-1 | X thread | F01 F03 F06 | 1/ I practiced an inbox workflow with 8 synthetic emails: define classification rules, then review a table, reply drafts and a daily brief. This shares the practice method; a real inbox workflow remains unverified. |
| X-2 | X thread | F02 | 2/ The rules use 6 business labels plus Review for uncertainty. Missing context becomes a question; each email has one primary workflow label. |
| X-3 | X thread | F03 F04 | 3/ There are 2 reply drafts. Unknown meeting times, prices and delivery dates stay pending for a person to decide. |
| X-4 | X thread | F05 | 4/ The sequence: calibrate samples → check real tools → small trial → acceptance → Skill → consider a Routine. Review each stage separately. |
| X-5 | X thread | F06 | 5/ This exercise measured no time savings or accuracy and wrote nothing to Gmail. Define inputs, outputs and exceptions before deciding whether to automate. |
| LI-01 | F01–F06 | I used 8 synthetic emails to practice defining a clear deliverable. The output is a classification table, 2 reply drafts and a daily brief. Six business labels plus Review retain emails with missing context. Two checks matter: each judgment traces to its source email, and a draft never confirms an unknown time, price or delivery date. Real Gmail tools, a small live trial and human acceptance still need checking before a Skill or Routine. No time savings or accuracy were measured, and no real inbox was written. Which recurring deliverable would you define first? |
These are reference texts without a real Typefully draft_id.
The JSON omits publish_at/plan_at and sets share:false; it was not executed.
Editorial review — synthetic sample Input:1 article, F01–F06. Output:5 X posts and 1 LinkedIn post. Facts: all numbers trace to the article; synthetic/unverified qualifications retained. Review needed: tone, account identity, links, platform length/format. Not done: live draft creation, scheduling, publication or performance measurement. Reject any invented efficiency percentage.
Prepare the materials and check your tools
Start with the practice files
- Download the source article and editorial brief.
- Choose channels and permitted tone examples.
- List fixed numbers, dates and qualifications.
Before using real services
- Verify social set and actual X/LinkedIn accounts.
- Check available draft/read/publish tools.
- Provide keys only through secure credential input, never prompt files.
Complete one file-based exercise first. For real services, verify the actual account, available tools and operation scope, then follow the acceptance check at each stage.
Download the original input and rules ↓Execution and takeover boundaries
The actions below require the agreed inputs, accounts and scope. Configuring a Routine does not expand that authorization.
Can run automatically
Prepare local channel-specific drafts and a fact map from the supplied article, without inventing performance data.
Requires confirmation
Confirm the social set, platform identity and text before saving to Typefully. Scheduling, sharing and publishing require separate approval.
Take over after a failure
After a save timeout, inspect the account and draft id before retrying and pause writes. Deliver files if saving is unavailable; have the user inspect unintended schedules.
Extract facts and retain qualifications
Use the article as the sole factual source. Separate sample events, recommendations and unmeasured outcomes.
Two drafts cannot become two sent emails, and synthetic practice cannot become a customer case.
Read source-article.md only. Create facts.md with F ids, source evidence, type(sample fact/advice/unverified), usable wording and required qualifications. Do not browse, imply a real Gmail test or invent time-saving/accuracy numbers.A traceable F01–F06 fact table.
The meanings of 8, 6 and 2 remain; F06 stays unmeasured.
Define channel structure, reader and tone
X can use one idea per post; LinkedIn can explain context and method in paragraphs.
Before later publication, check actual editor length, links and previews instead of hardcoding changing platform limits.
Use editorial-brief.json and facts.md to outline 5 X posts and 1 LinkedIn post with fact ids. State synthetic practice early. Show outlines only; do not connect accounts or schedule.Two different outlines with consistent facts.
Advice has not become accomplished work; each paragraph has a source id.
Generate drafts and audit facts separately
Review factual accuracy before repetitions, exaggeration, unexplained abbreviations and inconsistent voice.
The reference is one possible answer; structure can change but sample status cannot.
Generate drafts from the confirmed outlines with a draft-id→fact-id table. Independently check invented numbers, missing qualifications, plans written as done, repetition and unapproved promises. Revise and record changes in drafts.md only.5 X texts, 1 LinkedIn text and a fact map.
All claims trace to the article; both versions retain synthetic/unverified scope.
Inspect account identity and the draft request
The payload is an unexecuted request example. Verify current API fields and account capabilities before use.
Confirm social set and platform identities, not only display names.
Read-only inspect <Typefully connection/social set> and corresponding X/LinkedIn accounts/tools. Review draft-payload.json for ordinary drafts: no publish_at, plan_at or immediate publication parameters; share:false. Show checks/request only; do not write or echo keys.A target account list and request preview.
No scheduling/publication fields or public share request.
Save a specifically approved draft and read it back
After content/account confirmation, save one draft and inspect its actual draft_id, text, platform and state.
A timeout may follow successful creation. Check existing drafts before retrying.
Only after specific account/request approval, create one ordinary draft without scheduling, publishing or public sharing. Record draft_id and read back title, platforms, body and draft state. If tools fail, retain files marked unwritten. On uncertainty check duplicates; never overwrite existing edits.A real id/readback or explicit unwritten status.
Account/format correct; no paragraphs missing, schedule or published record.
Save the method and review publication separately
A new article needs a fresh fact list. Do not inherit the sample’s numbers or old account assumptions.
Keep timing advice as a document until the author approves a specific publication action.
Draft a content-repurpose Skill: source/brief intake, facts, outlines, rewriting, audit and draft/source delivery. Typefully writing, public sharing, scheduling and publishing are separate stages; default to local files. Create no timed publishing task or cross-Bot delegation.A reusable Skill draft and editorial checklist.
New input starts with facts and inherits no account/time/publication permission.
Find the evidence before choosing the next action
An efficiency percentage appeared
Return to the fact map; unmeasured stays unmeasured.
A post lacks context
Keep each post understandable and repeat necessary qualifications.
Saving failed or timed out
Check target drafts and actual draft_id before retrying.
The draft became scheduled
Stop writes and inspect fields/platform state with the user. Draft approval was not publication approval.
Platform preview differs
Check actual links, paragraphs, length and media; platform rendering is unverified here.
Check every item before finishing
These page-local checkboxes reset on refresh. Checking them does not verify a real service connection.
Checked 0 / 6 items
Download the inputs separately from the reference answers
Start with input and rules without showing the reference answers. After acceptance, save the report and complete prompt set.
source-article.md↓Editorial briefChannels, tone and factual constraints.editorial-brief.json↓Reference draftsFive X posts and one LinkedIn post.expected-drafts.md↓Ordinary draft request exampleNo keys, publishing or scheduling; not API-tested.draft-payload.json↓Draft-to-fact mapSix drafts linked to source ids.fact-map.csv↓All stage prompts6 instructions with acceptance checks.practice-prompts.txt↓Check documented capabilities and exercise results separately
Content updated: · Editing a page does not reverify its sources or live workflow.
Documentation review: Unverified · no review date recorded
No case-specific documentation review is recorded. References below are provided for review and do not establish verification.
Live workflow verification: Unverified
Live accounts, external writes, scheduled execution and long-term operation are unverified. Synthetic exercises and page checkboxes do not establish live verification.
The article and reference drafts are synthetic. No Typefully connection, draft creation, scheduling or publication was performed.
- Typefully: API overview ↗
Review authentication, supported platforms and draft workflows; no API call was made here.
- Typefully: API reference ↗
The request body is an unexecuted example. Check current fields and account capabilities before use.
- Grok Bot: apps and the shared computer ↗
Check connected accounts, available tools, visibility and user takeover.
- Grok Bot: Skills and Routines ↗
Validate a manual run before saving a method or scheduling it.
Try another small batch of your own material
Retain original input and operation records. Verify that each result can be traced before adding connections or recurring runs.