docs(mode-economics): refresh stale token table

[dpol1]

Summary

• docs/mode-economics.md is the reference adopters use to estimate token
  costs, but its tables had drifted behind the shipped skill set: the Pairing
  table still listed the multi-agent review pipeline as a "future skill", and
  five shipped skills were missing rows entirely.
• Rename the "Multi-agent review pipeline" row to pairing-multi-agent-review
  (shipped, experimental — range unchanged) and add the missing rows:
  pre-first-pr-check (Pairing) plus the four Mentoring-family skills the
  table omitted (mentoring-welcome, onboarding-concierge,
  contributor-to-committer, good-first-issue-sweep).
• New ranges are marked Estimated, following the table's existing
  convention for experimental skills.

Type of change

• Documentation (docs/, README.md, CONTRIBUTING.md)

Test plan

• prek run --files docs/mode-economics.md passes (markdownlint, typos,
  lychee link check all green)
• Docs-only change; no behaviour to test. Skill inventory cross-checked
  against docs/modes.md (Mentoring: 7 shipped, table had 3; Pairing: 3
  shipped, table had 1 + a "future" placeholder).

RFC-AI-0004 compliance

Docs-only; no principles touched.

Linked issues

Closes apache/magpie#881

Notes for reviewers

The four new Mentoring rows and pre-first-pr-check carry estimated ranges
anchored on analogous existing rows (newcomer-issue-explainer for the
single-thread drafters, issue-reassess for the backlog sweep). The two
worth sanity-checking are contributor-to-committer (15K–50K — activity
history read) and good-first-issue-sweep (20K–80K per 10-issue sweep).

[potiuk]

Approving — this is a clear improvement and the ranges are sensibly
conservative. Two things worth recording, neither blocking.

The refresh is partial

The summary frames this as catching the table up with the shipped skill set,
but four shipped skills still appear nowhere in mode-economics.md after this
change — all family: contributor-growth:

• contributor-activity-sweep
• committer-onboarding
• contributor-sentiment
• contributor-nomination

AGENTS.md § Keeping evals and mode-economics in sync treats a new skill as a
trigger to update this doc, so these are a real pre-existing gap rather than a
deliberate omission. Not this PR's job to close — the Mentoring/Pairing scope
here is coherent — but a follow-up would finish the thought.

Minor: the four added rows aren't all "Mentoring-family"

The PR description calls them "the four Mentoring-family skills", but only
mentoring-welcome and good-first-issue-sweep are family: mentoring;
onboarding-concierge and contributor-to-committer are
family: contributor-growth. Likewise pre-first-pr-check is
family: pr-management, not pairing.

The placement is fine — this page's sections are modes, not families, and the
table already carries pr-management-mentor under Mentoring — so no change is
needed in the diff. Flagging only so the description doesn't get quoted later as
evidence of a family mapping it doesn't establish.

---

This review was drafted by an AI-assisted tool and
confirmed by an Apache Magpie maintainer. The maintainer
approving this PR has read the findings and signed off. If
something feels off, please reply on the PR and a maintainer
will follow up.

More on how Apache Magpie handles maintainer review:
CONTRIBUTING.md § Opening a pull request.