[JarryShaw]

• 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
• Added a changelog entry — 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?

• ci — workflows or build tooling

---

Description of your pull request and other information

Follow-up to #2128 and the new docs/source/engines.rst page (f9de043d6, d7aa41535), so a benchmark report pastes onto the page without hand edits.

• examples/benchmark/report.py footnotes: it emits reST footnotes instead of literal † / ‡ / ¶ markers and hand-written legends. Each noted row or column gets [*]_, overlapping rows share a labelled [#overlap]_, and the bodies follow each table in reference order. A gap with no recorded reason now gets an explicit "no reason recorded" note instead of a dangling marker.
• report.py attribution: the "Measured with" line names the script that actually ran. The new --runner / --workflow options are fed by native.sh and, through entrypoint.sh, by run.sh. Every index.rst reference becomes engines.rst, in the script and in examples/benchmark/README.md.
• benchmark.yml: timeout-minutes drops from 300 to 90. The first full run, 40062653380 on ubuntu-26.04-arm with every engine, took about 38 minutes.
• Tests: TestFootnotes in examples/benchmark/test_harness.py parses the output with docutils, so each reference must reach its own row's note. Against main, 20 tests fail. A new test pins the timeout.
• Check: run 40062653380's output, spliced into engines.rst, changes no words. Only two tables' first-column padding differs.

[JarryShaw]

GOOD TO GO at 208047c7f (Sonnet review).

• Footnotes: a docutils render of run 40062653380's output, spliced into engines.rst, has no footnote, table or indentation errors. Each [*]_ reaches the body after its own table, alongside the page's existing footnotes. The shared [#overlap]_ resolves several references to one body, which the tests pin.
• Gaps and attribution: a gap with no recorded reason gets an honest "no reason recorded" note, not a dangling reference. The --runner / --workflow plumbing (native.sh; run.sh → entrypoint.sh via BENCH_RUNNER) is safe under set -euo pipefail and on bash 3.2.
• Timeout: 90 min is about 2.3× the measured 38 min on arm, and about 50 min is estimated for ubuntu-latest (1.31× slower per packet in run 40047570595). Pyshark has only been measured on Linux, so a Windows or macOS dispatch with pyshark is the one case that could run long.
• Tests: 171 passed. With main's examples/benchmark/, 20 tests fail.

[github-actions]

Coverage: 91.55% (unit tier, Python 3.15, 208047c7f, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	2	95.16%
pcapkit/const	19599	880	2618	866	92.12%
pcapkit/corekit	2589	75	816	37	96.48%
pcapkit/dumpkit	282	3	102	3	98.44%
pcapkit/foundation	2971	23	1036	25	98.80%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17418	219	4574	185	98.14%
pcapkit/toolkit	869	36	298	10	95.20%
pcapkit/utilities	498	7	144	5	98.13%
pcapkit/vendor	4432	1988	1018	167	51.87%

Per-file detail: the coverage-html artifact of this run.

[JarryShaw]

Ready to merge at 208047c7f.

• Review: the Sonnet review is GOOD TO GO.
• CI: every blocking check passed (Required checks passed SUCCESS). The 3.8–3.10 KTLO legs carry no warnings.
• Ordering: the tests/project plain-unittest ordering leg passed on a scratch merge into main d7aa41535: 2233 tests, 0 failures.
• After merge: the next Engine Benchmark report pastes into docs/source/engines.rst without hand edits, apart from that page's own intro paragraph and Run row. A Windows or macOS dispatch with pyshark has never been timed against the new 90-minute limit.