# Phase 7 brief: scoring Cargo-Planner on its public documentation

This is the exact brief the phase 7 auditor received. It is kept as part of
the methodology disclosure (see `00-methodology.md`).

## Your role

You audit **Cargo-Planner** (cargo-planner.com) as a ninth product, to exactly
the same standard as the eight competitors before it. Cargo-Planner is also the
company that commissioned this audit. That makes errors in its favour the
worst kind of error this audit can make. Where you are unsure, resolve the
doubt the way you would for any other vendor, and flag it.

## Rules

Everything in `02-audit-brief.md` and `04-scoring-brief.md` applies, with one
change: you are required to read cargo-planner.com, which those briefs
forbade.

- **Buyer emulation.** Use only what a prospective buyer can reach on the live
  public site: documentation, guides, feature pages, blog, API documentation,
  release notes and downloadable files. Do not log in, start a trial or use
  video.
- **The live site is the evidence.** Do not read the website's source code or
  any other file in `/workspaces/cplweb` except the following, all in
  `research/market-audit/`:
  - `00-methodology.md`, `02-audit-brief.md`, `04-scoring-brief.md`, this brief
  - `03-taxonomy.json`, `06-rules.json`
  - your own output files

  Specifically, do not read `work/`, `src/`, the other products' audit or
  score files, or any page source. A buyer sees the published site, not the
  repository.
- **Marketing pages are marketing.** Feature, solution, comparison and
  capability pages are vendor marketing. A statement there counts as
  `documented` only if it explains how the capability works in enough detail
  to meet the pass condition, exactly as for competitors. Pages that compare
  Cargo-Planner with competitors are not evidence about competitors, and do
  not change how you score Cargo-Planner.
- **Quote, don't paraphrase.** Every cell carries verbatim quotes with the
  URL.

## Steps

1. **Open reading.** Read the documentation broadly first, as the phase 2
   auditors did. List the sources you read. List the product's general
   mechanisms (rule engines, custom attributes, configurable objectives),
   saying what each one can select on and what it can do.
2. **Scoring.** Score all 149 rows of `03-taxonomy.json` (draft-2), recording
   each cell exactly as `04-scoring-brief.md` specifies: evidence, mechanism,
   availability, quotes, recipe, searched, variant, note, ambiguous.
   - A composable cell needs a recipe with a citation for every step. Say
     whether the site itself shows that combination for that purpose
     (`vendor-shown`) or whether you assembled it (`auditor-built`).
   - A `not-documented` cell needs a `searched` entry.
3. **Issue types.** For every cell that is ambiguous, composable, a
   workaround, claimed, beta or announced, add a `categories` array using the
   issue type ids defined in `06-rules.json` (`categoryDefinitions`):
   - `composable-vendor-shown`, `composable-auditor-built`, `workaround`
   - `screenshot-only`, `undescribed-field`, `enforcement-unclear`
   - `implied-not-stated`, `concedes-gap`, `shared-mechanism`
   - `restricted-scope`, `source-conflict`, `outdated-source`
   - `marketing-only`, `not-generally-available`, `other`


## Output

Write `/workspaces/cplweb/research/market-audit/scores/cargo-planner.json`.
Use the same shape as the other score files, plus `sources`,
`generalMechanisms` and a `categories` array in each cell. Save after each
cluster. Check with a script that there is exactly one cell per row.

Your final reply should give:

- the counts by evidence, by mechanism and by availability;
- the number of ambiguous cells;
- the rows where you are least sure;
- the places where the site's marketing claims more than its documentation
  supports.