[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 — changes to CI configuration files and scripts

---

Description of your pull request and other information

The fourth create-release batch for #1967, covering the conda job and pypi's KTLO summary. Each move meets the #1967 inline rule: the shell is duplicated or has real logic.

• util/release_conda.py leg|build|upload: conda's already-on-Anaconda check, plus Build conda packages and Upload conda packages to Anaconda (Intel). The last two are shared by create-release's conda and cron-conda's conda-dist, which drops the duplicated shell and both case maps.
• pypi: KTLO summary now runs the shared leg_summary.py ktlo. Record this KTLO leg's outcome, 6 lines, stays inline under the rule.
• Kept inline: two-command steps, and all of github and tag, which come in the next batch. Get Version stays inline too.
• Equivalence: test_release_conda_scripts_1967 keeps the old bodies verbatim and runs each beside its script, requiring the same status, stdout, stderr, outputs and calls. Six planted divergences were caught, and a 2xx other than 200 is pinned as not published, as the old shell treats it. actionlint 1.7.12 is clean.
• Unavoidable differences: labels are whitespace-collapsed rather than xargs-trimmed, bash's line N: prefixes are gone, and packages uses the helper's heredoc delimiter.
• Windows (untested on a runner): the build now gets native Windows paths, not Git Bash /tmp ones. That leg is non-blocking and first runs on Saturday's schedule (Conda: bring back the Windows (win-64) package, in two phases #2060).

[JarryShaw]

GOOD TO GO at df521c91e (Sonnet review).

• Workflow diff: a PyYAML diff of both workflows shows only run: bodies changed, plus the new PCAPKIT_CONDA_OS/PCAPKIT_PYTHON_VERSION env entries. The outputs and their order match.
• Equivalence test: five of its six embedded old bodies are byte-identical to main. The sixth comes from the cron-conda build and differs only in one comment line and the Windows arm, which it strips. The test caught 7 of 8 planted divergences; the miss was a 2xx other than 200 counting as published, a small follow-up.
• Windows leg: UNVERIFIED until Saturday's run. No break was found: conda.exe should resolve from the activated PATH, and the native temp path is what the smoke step's Windows Python needs. The leg stays non-blocking.
• Scripts: stdlib only and 3.7-parseable. Eleven test files pass.

Possible follow-up from review: put the 6-line Record this KTLO leg's outcome back inline, under a "trivial shell stays inline" rule. That is pending.

[JarryShaw]

GOOD TO GO at 34a58cf1d (Sonnet delta review of bad2e8522, plus a whitespace-only amend).

• KTLO step: a PyYAML compare shows Record this KTLO leg's outcome is identical to main (env and run). release_ktlo appears nowhere (git grep).
• 2xx check: the old shell reads the reply only on an exact 200, and release_conda.py matches. Cases for 201 and 204 run against the old shell, and test_only_a_200_is_read pins it.
• Inline rule (ci: divide and simplify the workflows into scripts, shared heads and smaller files #1967): every remaining move qualifies. The three conda steps are 48 lines with a bug history, or are duplicated with cron-conda (50/55 and 28/28 lines). pypi's KTLO summary reuses leg_summary.py ktlo, which four other jobs already call.
• Amend: 34a58cf1d only fixes two blank-line spacings (pycodestyle E302/E303) in test_release_conda_scripts_1967.py.
• Tests: five test files pass; the file I re-ran after the amend passes too.

[github-actions]

Coverage: 91.54% (unit tier, Python 3.15, 34a58cf1d, 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	2565	71	812	37	96.56%
pcapkit/dumpkit	282	3	102	3	98.44%
pcapkit/foundation	2930	23	1022	25	98.79%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17414	219	4574	185	98.14%
pcapkit/toolkit	869	36	298	10	95.20%
pcapkit/utilities	473	7	136	5	98.03%
pcapkit/vendor	4432	1988	1018	167	51.87%

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

[JarryShaw]

Ready to merge at 34a58cf1d.

• Review and CI: the Sonnet delta review is GOOD TO GO, and every blocking check passed (Required checks passed SUCCESS). The 3.8–3.10 KTLO legs carry no warnings.
• Merge into current main: the branch is one test-only commit behind (403117c98, test: a GraalPy xdist worker died during a SIGALRM TimeLimitTests case #2084), so I merged it into main in a scratch tree. That merge was clean, and the plain-unittest ordering leg for tests/project ran on it: 2117 tests, 0 failures.