verify_report_figures
shallowio.github.valuein/mcp-sec-edgar · Verify this server
Check EVERY number in one of your own reports against the filing it cites, in one call. Returns a verdict per figure: `matches` (the prose quotes the filing's own rendered string), `differs` (positive evidence the filing says something else — the report is wrong), or `not_checkable` (no comparison was possible). CALL THIS BEFORE `publish_report` OR `sign_off_report`. `publish_report` REFUSES a report carrying an unreviewed `differs`, so running this first is how you find out what to fix instead of being refused. TO FIX A `differs`: rewrite the figure by quoting `verdict.display` character-for-character (or `verdict.derived_quarterly_display` if you meant the quarter rather than the year-to-date number) with `update_report`, then call this again. A non-empty `verdict.scope` means the citation points at the WRONG FILING — re-quoting will not fix that; correct the citation. If the report's number is right and the comparison is not (an external source, a deliberate restatement), record a human disposition with `save_figure_review` — a reviewed figure no longer blocks publishing. ⚠️ `not_checkable` IS NOT A PASS AND IS NOT A RETRY. It means nobody compared anything — a figure with no fact, a fact with no rendered string, a lookup that failed. Report it as unverified; do not call the tool again expecting a different answer, and never present it as verified. Expect it to be a large share of any real report. `untraced` is counted separately from `not_checkable` on purpose: a figure that never had a source is a different finding from one whose source could not be read. Nothing here is stored — a verdict is recomputed every call, because a later filing can restate a number and an edit changes a figure's identity. Only the report's author can verify it; a report that is not yours is indistinguishable from one that does not exist. Tier: sp500+ (sample rejected). Free — provenance calls are never charged.
1 trials · measured 33 days ago
verify_report_figures scores 100.0/100 on Vouch's measured behaviour index, from 1 real invocation trials against io.github.valuein/mcp-sec-edgar, measured 16 Sept 2026 under methodology v0.2.0. Every measured component scored 100.
Component breakdown
| Component | Weight | Value |
|---|---|---|
| Reliability | 35% | not applicable |
| Schema integrity | 25% | 100.0 |
| Failure behaviour | 15% | not applicable |
| Latency | 15% | not applicable |
| Concurrency | 10% | not applicable |
Tool details
- Transport
- remote
- Credential class
- self-provisionable
- Category
- Finance & compliance
- Input schema
- not declared
- Output schema
- not declared
- Side-effect classification
- unclassified
Score history
| Day | Score | Tier | Methodology |
|---|---|---|---|
| 2026-09-16 | 100.0 | shallow | v0.2.0 |
Probe evidence
| Probe | Outcomes |
|---|---|
| schema_integrity | pass: 1 |
Raw request/response logs are not archived yet — the outcome counts above are drawn directly from every recorded trial.
Embed this score
Available for every tool, scored or not — not a verification perk. Always links back to this page.
[](https://vouch.tools/tools/5865b05e-5c09-46cb-ad95-39b6160ff196)