docs(toolkit): drop agreement claims from adapter docstrings (#1708)

[JarryShaw]

Please follow the guide below

• You will be asked some questions, please read them carefully and answer honestly

• Put an x into all the boxes [ ] relevant to your pull request (like that [x])

• Use Preview tab to see how your pull request will actually look like

• Searched for similar pull requests

• Followed the coding style (make pylint, make mypy, make isort)

• make test passes, and a test case covers the change. Narrow runs only, listed below.

• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible: N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

What is the purpose of your pull request?

Tick the commit type your subject line carries.

• fix — corrects a defect
• feat — adds a feature
• perf — changes performance, not behaviour
• refactor — changes neither behaviour nor performance
• test — tests only
• docs — documentation only
• ci — workflows or build tooling
• release — bumps the version or rolls up a distribution
• chore — anything else

---

Refs #1708 (phase 3b). This PR removes the agreement claims from the adapter docstrings, following the best-effort ruling on #1687. The docs pages were fixed in d06701064.

• It rewrites or drops 14 agreement clauses across foundation/engines/{dpkt,pypcapfile,pyshark}.py, foundation/extraction.py, toolkit/{dpkt,pyshark}.py and foundation/reassembly/ip.py. Every replacement claim was checked against the code, for example the Loopback families {2, 24, 28, 30} and localcontext(prec=64).
• The change is docstring- and comment-only: with docstrings stripped, the AST of all seven modules is identical to main. Four paragraphs are reflowed (whitespace only).
• Guard: tests/project/test_no_engine_agreement_prose_1708_unit.py scans pcapkit/**/*.py and docs/**, with reST markup stripped first. It fails on main with 19 hits and passes here. Legitimate uses such as pypcap-parity and the pcap-ct comparisons stay clean. Two hits sit in files the open feat(utilities): one config module for the tunable limits (PCAPKIT_* environment variables), and one home per duplicated protocol constant #1726 work owns and are allowlisted with that reason.
• Sphinx warnings are the same set before and after.

[JarryShaw]

Verdict: GOOD TO GO at c69265ad3. A Sonnet cross-review checked it (author: Opus).

• Code unchanged: with docstrings stripped, the AST of all seven modules matches main. Checked twice: once removing only docstrings, once every bare-string statement. The reflowed paragraphs keep exactly the same words; only the line breaks moved.
• Guard: fails on main with 19 hits and passes here. It caught all six issue shapes, plus the default-markup comment. Thirteen near-miss phrasings of the reviewer's own raised no false positive.
• Replacement claims match the code: the Loopback families and prec=64.
• Sphinx: no warnings added or removed. test_isort_clean and test_documentation_claims pass.
• My own re-run in the PR's tree: 5 passed for the guard and test_isort_clean.
• Left for feat(utilities): one config module for the tunable limits (PCAPKIT_* environment variables), and one home per duplicated protocol constant #1726: toolkit/pypcapfile.py:206-207 still says a MAC address takes "the same form ... for the default engine". It sits in a file the feat(utilities): one config module for the tunable limits (PCAPKIT_* environment variables), and one home per duplicated protocol constant #1726 work owns, so refactor(toolkit): revert the cross-engine agreement tests and the adapter code that only mirrors the default engine #1708 stays open until that line is fixed.