Browse docs

Configured tool · read-only

get_digest

Read the latest completed workspace week as descriptive evidence.

read-onlynon-destructiveclosed-world

Human and agent roles

Configured boundary

Configured workspace

This tool reports what happened. It does not recommend what to publish, and it never ranks or aggregates a null metric as zero.

This reference describes a connection made with @postdom/mcp@0.4.0 and a workspace-scoped agent key created in Accounts.

Callable schema

Input fields

Field names match the configured tool schema. Runtime-only constraints are labeled separately from schema validation.

FieldTypeContract

Synthetic values

Example call

Identifiers and values are examples. They do not represent a connected workspace or accepted media source.

EXAMPLE CALL · CONFIGURED TOOL SCHEMA
get_digest({})

Observed API record

Return shape

Returns the latest append-only digest version, completed period, generation time, and structured evidence content. The summary and the activity counts are whole; the evidence lists behind them are bounded to five entries each, and each carries a total. Before the first completed week it returns digest_not_found.

version · period · generated_atidentity
Immutable digest version and exact completed-week bounds.
content.summarystring
The week in one sentence, whole.
content.population_definitionstring
Which posts the week's figures were counted over, stated once for every figure below.
content.activitycontextual figures
Recorded, published, failed, partial, human-rejected, and other-state counts, each with its observed and total populations and its polling state. Excluded is total minus observed.
content.runs · content.runs_totalrun evidence[] · number
Up to five observed posts with their exact state and recorded failure reasons, and the count of every run in the week.
content.top_posts · content.top_posts_totalevidence ranking[] · number
Up to five ranked only where the chosen metric is evidence-backed, and the count of every ranked entry.
content.failure_reasons · content.failure_reasons_totalreason evidence[] · number
Up to five recorded failure reasons with their counts, and how many distinct reasons the week held.
content.coverage_notes · content.coverage_notes_totalstring[] · number
Up to five polling cadence endings and availability gaps, and how many the week held.
content.week_over_week_accountsnumber
How many accounts have a completed week-over-week comparison. Read the comparison itself with get_performance on the account.
truncatedflag | optional
Present when any list above was bounded. Absent means every list is complete.

Stop conditions

Agent handoff

available

Read before planning

Preserve every period, population, polling state, and coverage note you are given. Use it only as evidence of what happened.

truncated

Say the lists are partial

Report each list against its total. A bounded list is evidence behind a count, never the whole of it.

digest null

Continue without inference

No completed weekly digest exists yet. Do not invent a learning or recommendation.

VERIFIED · 26 AUGUST 2026GET /v1/digest · learning_digests

Examples are synthetic. Human workspace-key creation and revocation remain in the authenticated Accounts screen.