Verification trail
#35
✅ Verified
.claude/commands/docs-review/scripts/testdata/v3-fixture-author.md.txt:94
Full mechanics: [CONTRIBUTING.md §AI-assisted contributions](https://github.com/pulumi/docs/blob/master/CONTRIBUTING.md#ai-assisted-contributions).
evidence: The repo's CONTRIBUTING.md has the heading "## AI-assisted contributions" at L17, so the #ai-assisted-contributions anchor resolves. The section starts: "The repository runs a tiered review pipeline on every PR. AI-assisted contributors…; source: repo:CONTRIBUTING.md L17
source: repo:CONTRIBUTING.md L17
framing: exact-match
pass1 · url · high confidence
#1
➖ Not a claim
.claude/commands/docs-review/references/output-format.md:479
**Full evidence:** [verification trail, …](%%EVIDENCE_URL%%). ← no emoji (cards before 2026-09-25 led with 📎; readers accept both)
framing: The regex matched a date in an internal template comment. It is the author's own design, not a claim that can be checked.; evidence: This line is a template in the docs-review tool's own output-format spec. It describes the author's own card format: the evidence link no longer uses an emoji, and cards from before 2026-09-25 used 📎. It describes the PR author's own…; source: .claude/commands/docs-review/references/output-format.md:L479 (internal tooling spec; regex matched the date)
source: .claude/commands/docs-review/references/output-format.md:L479 (internal tooling spec; regex matched the date)
framing: The regex matched a date in an internal template comment. It is the author's own design, not a claim that can be checked.
pass3 · numerical · high confidence
#2
➖ Not a claim
.claude/commands/docs-review/references/output-format.md:492
the anchored line. So quote the source text verbatim, and put only its replacement in the
framing: The source_hint "So" is a conjunction the regex matched, not a cited source.; evidence: This line is an instruction in the PR author's own review-output format spec. It tells reviewers to quote the source text word for word and put only its replacement in the suggestion. It doesn't state a fact about any third party, so…; source: .claude/commands/docs-review/references/output-format.md L492 (claim text itself; internal tooling instruction)
source: .claude/commands/docs-review/references/output-format.md L492 (claim text itself; internal tooling instruction)
framing: The source_hint "So" is a conjunction the regex matched, not a cited source.
pass3 · attribution · high confidence
#3
➖ Not a claim
.claude/commands/docs-review/references/output-format.md:500
row under, and the update lane restores the pair (`ensure_author_sections`) before placing a
framing: Describes the author's own design, and cites no source.; evidence: (escalated from pass3) L500 is part of the PR author's description of their own docs-review pipeline: "the update lane restores the pair (`ensure_author_sections`) before placing a reopened, added, or promoted row." It names an internal…; source: repo:.claude/commands/docs-review/references/output-format.md L495-504; gh search code --owner pulumi ensure_author_sections (no results)
source: repo:.claude/commands/docs-review/references/output-format.md L495-504; gh search code --owner pulumi ensure_author_sections (no results)
framing: Describes the author's own design, and cites no source.
pass1 · entity-spec · medium confidence
#4
➖ Not a claim
.claude/commands/docs-review/scripts/ack-request.sh:49
case "$1" in
framing: The regex false-positive read the shell positional parameter "$1" as a price.; evidence: The flagged text `case "$1" in` is a line of shell syntax from a helper script (ack-request.sh). It starts a case statement on the first positional argument and makes no factual claim. The regex likely flagged `$1` as a dollar amount.; source: .claude/commands/docs-review/scripts/ack-request.sh:L49
source: .claude/commands/docs-review/scripts/ack-request.sh:L49
framing: The regex false-positive read the shell positional parameter "$1" as a price.
pass3 · numerical · high confidence
#5
➖ Not a claim
.claude/commands/docs-review/scripts/ack-request.sh:52
--message) MESSAGE="$2"; shift 2 ;;
framing: Code, not prose: the author's own tooling logic, with no third-party assertion.; evidence: This line is a bash case branch in the PR author's own helper script (`ack-request.sh`). It parses a `--message` argument and shifts past two positional args. The regex likely flagged `shift 2` as a number, but the line makes no factual…; source: .claude/commands/docs-review/scripts/ack-request.sh:L52-55
source: .claude/commands/docs-review/scripts/ack-request.sh:L52-55
framing: Code, not prose: the author's own tooling logic, with no third-party assertion.
pass3 · numerical · high confidence
#6
➖ Not a claim
.claude/commands/docs-review/scripts/ack-request.sh:56
*) echo "ack-request.sh: unknown argument: $1" >&2; exit 2 ;;
framing: Shell code in the author's own tooling; 'exit 2' is an exit status, not a numerical assertion.; evidence: This line is the default branch of a case statement in the PR author's own helper script (it prints an unknown-argument error and exits with status 2). It describes the script's own behavior, not a third-party fact that could be checked.…; source: .claude/commands/docs-review/scripts/ack-request.sh:L56
source: .claude/commands/docs-review/scripts/ack-request.sh:L56
framing: Shell code in the author's own tooling; 'exit 2' is an exit status, not a numerical assertion.
pass3 · numerical · high confidence
#7
➖ Not a claim
.claude/commands/docs-review/scripts/build-evidence.py:552
here" (reader feedback, 2026-09-25). The NOTE is the whole message.
framing: The date in the comment is when the feedback was recorded, not a checkable numerical claim.; evidence: The line is a code comment in the PR's own tooling script (build-evidence.py). It records internal reader feedback dated 2026-09-25 and explains the author's design ("The NOTE is the whole message"). There's no external or numerical fact…; source: .claude/commands/docs-review/scripts/build-evidence.py:L552
source: .claude/commands/docs-review/scripts/build-evidence.py:L552
framing: The date in the comment is when the feedback was recorded, not a checkable numerical claim.
pass3 · numerical · high confidence
#8
➖ Not a claim
.claude/commands/docs-review/scripts/compose-review.py:139
# cell gets a short excerpt (reader feedback 2026-09-25: the duplicate made
framing: A code comment describing the author's own pipeline; there is no third-party-attributed assertion to check.; evidence: This line is a code comment in the PR's own review-tooling script (compose-review.py). It records an internal design decision and the date of the reader feedback behind it, so there is no external fact to check. The regex flagged the…; source: .claude/commands/docs-review/scripts/compose-review.py:L139
source: .claude/commands/docs-review/scripts/compose-review.py:L139
framing: A code comment describing the author's own pipeline; there is no third-party-attributed assertion to check.
pass3 · numerical · high confidence
#9
➖ Not a claim
.claude/commands/docs-review/scripts/compose-review.py:1819
# until 2026-09-25; cards published before then keep it until they refresh,
framing: A date in a comment describing the author's own pipeline's transition behavior; not a falsifiable numerical claim.; evidence: This is a code comment in the PR author's own review-tooling script (compose-review.py). It describes an internal compatibility window for the tool's own review cards ("until 2026-09-25; cards published before then keep it until they…; source: .claude/commands/docs-review/scripts/compose-review.py:L1819 (internal tooling comment; no external source applicable)
source: .claude/commands/docs-review/scripts/compose-review.py:L1819 (internal tooling comment; no external source applicable)
framing: A date in a comment describing the author's own pipeline's transition behavior; not a falsifiable numerical claim.
pass3 · numerical · medium confidence
#10
➖ Not a claim
.claude/commands/docs-review/scripts/compose-review.py:1830
feedback, 2026-09-25). The H3 stays outside the fold: it is the section
framing: The regex matched the date '2026-09-25' in an internal code comment. It isn't a numerical claim about a product or a third party.; evidence: This line is a code comment in the PR author's own review-tooling script (compose-review.py). It gives the date of internal feedback (2026-09-25) and explains a layout choice about the H3 heading. It describes the author's own pipeline…; source: .claude/commands/docs-review/scripts/compose-review.py:L1830
source: .claude/commands/docs-review/scripts/compose-review.py:L1830
framing: The regex matched the date '2026-09-25' in an internal code comment. It isn't a numerical claim about a product or a third party.
pass3 · numerical · high confidence
#11
➖ Not a claim
.claude/commands/docs-review/scripts/compose-review.py:2084
before this PR can merge" (2026-09-01 update-lane smoke).
framing: The regex extractor flagged the date inside a comment about internal tooling. It is not a factual claim in published docs.; evidence: This text is a code comment in the PR author's own review-tooling script (compose-review.py). It points to an internal smoke-test run ("2026-09-01 update-lane smoke") that motivated a string in the pipeline. It describes the author's own…; source: .claude/commands/docs-review/scripts/compose-review.py:L2084
source: .claude/commands/docs-review/scripts/compose-review.py:L2084
framing: The regex extractor flagged the date inside a comment about internal tooling. It is not a factual claim in published docs.
pass3 · numerical · high confidence
#12
➖ Not a claim
.claude/commands/docs-review/scripts/compose-review.py:2087
because the full **How to answer** footer is collapsed (2026-09-25:
framing: The regex picked this up as a number because of the date stamp "2026-09-25". It is a note inside the author's own code, not a claim about how many of…; evidence: The line is a code comment in the PR author's own review-tooling script (compose-review.py). It describes how the author's pipeline collapses the "How to answer" footer, and the date is just the comment's change-log stamp. It says…; source: .claude/commands/docs-review/scripts/compose-review.py:L2087 (code comment; describes the author's own design)
source: .claude/commands/docs-review/scripts/compose-review.py:L2087 (code comment; describes the author's own design)
framing: The regex picked this up as a number because of the date stamp "2026-09-25". It is a note inside the author's own code, not a claim about how many of something there are.
pass3 · numerical · high confidence
#13
➖ Not a claim
.claude/commands/docs-review/scripts/compose-review.py:2128
"- **Why:** <TODO: 1-2 sentences — what is wrong (🚨) or what only the author can settle (❓)"
framing: The regex flagged "1-2" as a number. It is really a length guideline inside the author's own template text.; evidence: This line is a TODO placeholder in the PR author's own review-composition script. The "1-2 sentences" is a formatting instruction for the review template, not a factual number that could be checked against anything.; source: .claude/commands/docs-review/scripts/compose-review.py:L2128
source: .claude/commands/docs-review/scripts/compose-review.py:L2128
framing: The regex flagged "1-2" as a number. It is really a length guideline inside the author's own template text.
pass3 · numerical · high confidence
#14
➖ Not a claim
.claude/commands/docs-review/scripts/post-style-suggestions.py:80
leave the Files view (2026-09-25). It posts under its own marker, so the two
framing: The regex matched the date only. The line describes the author's own internal tooling design.; evidence: This line is a code comment in the PR author's own review-tooling script (.claude/commands/docs-review/scripts/post-style-suggestions.py). It describes the script's own posting behavior, uses its own marker, and gives a changelog-style…; source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L80
source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L80
framing: The regex matched the date only. The line describes the author's own internal tooling design.
pass3 · numerical · high confidence
#15
➖ Not a claim
.claude/commands/docs-review/scripts/post-style-suggestions.py:156
"_Generated by [Claude Code](https://claude.ai/code)_"
framing: A regex matched the URL inside a boilerplate "generated by" footer, not inside an assertion in the docs.; evidence: This line is a hard-coded attribution footer ("Generated by Claude Code") that the PR author's own docs-review script adds to the comments it posts. It describes the author's own tooling, not a third-party fact that could be checked. The…; source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L156 (script output template)
source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L156 (script output template)
framing: A regex matched the URL inside a boilerplate "generated by" footer, not inside an assertion in the docs.
pass3 · url · high confidence
#16
➖ Not a claim
.claude/commands/docs-review/scripts/post-style-suggestions.py:767
# Whole-quote rewrites in the same card ran 87-95% of the quote's length.
evidence: (escalated from pass3) This is a code comment in the author's own review script. It records something the author saw in their own tooling (whole-quote rewrites in a pulumi/docs#21645 review card) to explain why the FIX_PARTIAL_MIN_RATIO…; source: repo:.claude/commands/docs-review/scripts/post-style-suggestions.py L764-769
source: repo:.claude/commands/docs-review/scripts/post-style-suggestions.py L764-769
pass1 · numerical · high confidence
#17
➖ Not a claim
.claude/commands/docs-review/scripts/post-style-suggestions.py:780
# 43-156 chars; a one-sentence fence against a two-sentence line shares ~none.
framing: Internal code comment about the author's own heuristic; the regex flagged it only because it contains numbers.; evidence: This is a code comment in the PR author's own review-tooling script (post-style-suggestions.py). It explains the script's own string-similarity heuristic (typical line lengths of 43-156 chars, and little overlap between one-sentence and…; source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L780
source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L780
framing: Internal code comment about the author's own heuristic; the regex flagged it only because it contains numbers.
pass3 · numerical · medium confidence
#18
➖ Not a claim
.claude/commands/docs-review/scripts/post-style-suggestions.py:788
posted (2026-09-25 review of this mode) — a wrong button is worse than none."""
framing: Code comment describing the author's own design reasoning; the date is internal metadata, not a third-party assertion.; evidence: This is a line from a docstring in the PR author's own tooling script (post-style-suggestions.py). It explains an internal design decision and dates it to a review of that mode on 2026-09-25, which is today. It makes no outside factual…; source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L788
source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L788
framing: Code comment describing the author's own design reasoning; the date is internal metadata, not a third-party assertion.
pass3 · numerical · high confidence
#19
➖ Not a claim
.claude/commands/docs-review/scripts/post-style-suggestions.py:884
Shape rules, from 18 live Do-this blocks sampled 2026-09-25: the fence
framing: The regex flagged this as a numerical claim because of the sample count and date. It is actually an internal design note in a comment.; evidence: This is a code comment in the PR author's own review-tooling script. It describes how the author worked out the script's formatting rules from an internal sample of 18 of their own tool's "Do-this" output blocks. No public source could…; source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L884 (code comment; author's own pipeline description)
source: .claude/commands/docs-review/scripts/post-style-suggestions.py:L884 (code comment; author's own pipeline description)
framing: The regex flagged this as a numerical claim because of the sample count and date. It is actually an internal design note in a comment.
pass3 · numerical · medium confidence
#20
➖ Not a claim
.claude/commands/docs-review/scripts/test_apply_update.py:246
new_body = "x\n**Full evidence:** [verification trail](https://old.example/1).\n"
framing: Test data containing a dummy .example URL, not a cited source.; evidence: This line is a Python test-fixture string in a unit test (test_apply_update.py). It uses a placeholder URL on the reserved-style example domain (old.example), so it isn't a real citation. It doesn't make any assertion that could be checked.; source: .claude/commands/docs-review/scripts/test_apply_update.py:L246
source: .claude/commands/docs-review/scripts/test_apply_update.py:L246
framing: Test data containing a dummy .example URL, not a cited source.
pass3 · url · high confidence
#21
➖ Not a claim
.claude/commands/docs-review/scripts/test_apply_update.py:247
"x\n**Full evidence:** [verification trail](https://new.example/2).\n"
framing: Test fixture data using the reserved .example TLD; not a citation.; evidence: This is a test-fixture string in a Python unit test (test_apply_update.py). It uses the placeholder domain new.example (reserved under RFC 2606/6761) as a fake URL, so it makes no factual assertion that can be checked.; source: .claude/commands/docs-review/scripts/test_apply_update.py:L247-248
source: .claude/commands/docs-review/scripts/test_apply_update.py:L247-248
framing: Test fixture data using the reserved .example TLD; not a citation.
pass3 · url · high confidence
#22
➖ Not a claim
.claude/commands/docs-review/scripts/test_apply_update.py:283
"""Reader feedback 2026-09-25: resolved rows are history, so they fold
framing: The regex extractor probably flagged this because of the date "2026-09-25". It is a code-comment date stamp, not a numerical claim.; evidence: The line is a Python test docstring in the PR author's own docs-review tooling (test_apply_update.py). It records when internal reader feedback came in and how the author's pipeline handles it ("resolved rows are history, so they fold").…; source: .claude/commands/docs-review/scripts/test_apply_update.py:L283
source: .claude/commands/docs-review/scripts/test_apply_update.py:L283
framing: The regex extractor probably flagged this because of the date "2026-09-25". It is a code-comment date stamp, not a numerical claim.
pass3 · numerical · high confidence
#23
➖ Not a claim
.claude/commands/docs-review/scripts/test_apply_update.py:460
"""Reader feedback 2026-09-25: a nothing-for-you card spent eight lines
framing: The regex matched a date and the word "eight" in a code comment or docstring about the tool's own behavior.; evidence: This line is a docstring in the PR's own test script (test_apply_update.py). It records internal reader feedback about the review tool's output, where a card "spent eight lines". It describes the author's own pipeline, not a third-party…; source: .claude/commands/docs-review/scripts/test_apply_update.py:L460
source: .claude/commands/docs-review/scripts/test_apply_update.py:L460
framing: The regex matched a date and the word "eight" in a code comment or docstring about the tool's own behavior.
pass3 · numerical · high confidence
#24
➖ Not a claim
.claude/commands/docs-review/scripts/test_apply_update.py:501
"""Review finding 2026-09-25: the unprefixed `**Full evidence:**` line
evidence: This line is a test docstring in the PR's own tooling script. It describes an internal review finding dated 2026-09-25 about the `**Full evidence:**` line. It's the author's note about their own pipeline, not a checkable numerical or…; source: .claude/commands/docs-review/scripts/test_apply_update.py:L501
source: .claude/commands/docs-review/scripts/test_apply_update.py:L501
pass3 · numerical · high confidence
#25
➖ Not a claim
.claude/commands/docs-review/scripts/test_apply_update.py:515
"""Review finding 2026-09-25: the model writes the optional key as "",
framing: The regex flagged the date string as a numerical claim, but it is a code-comment provenance note.; evidence: This line is a docstring in the PR author's own test script (test_apply_update.py). It records an internal review finding dated 2026-09-25 about how the model writes an optional key. The regex picked up the date, but it is internal…; source: .claude/commands/docs-review/scripts/test_apply_update.py:L515
source: .claude/commands/docs-review/scripts/test_apply_update.py:L515
framing: The regex flagged the date string as a numerical claim, but it is a code-comment provenance note.
pass3 · numerical · high confidence
#26
➖ Not a claim
.claude/commands/docs-review/scripts/test_compose_v3.py:700
LONG_CLAIM = ("The GitHub repository at https://github.com/pulumi/examples/tree/master/"
framing: The regex picked up a URL inside the sample string of a test fixture.; evidence: The line is a string literal assigned to a test fixture constant (`LONG_CLAIM = (...)`) in a Python unit test for the docs-review tooling. It is sample input data for the test, not a factual assertion in published content.; source: repo:.claude/commands/docs-review/scripts/test_compose_v3.py L700
source: repo:.claude/commands/docs-review/scripts/test_compose_v3.py L700
framing: The regex picked up a URL inside the sample string of a test fixture.
pass1 · url · high confidence
#27
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:589
out, _ = pss.annotate_text(body, [{"file": path, "line": n}], files_url="https://x/pull/1/files")
framing: Test-fixture placeholder URL (host "x"); nothing to verify.; evidence: This line comes from a Python unit test. "https://x/pull/1/files" is a placeholder URL passed as a test argument to annotate_text, not a citation or factual assertion. The regex URL extractor flagged it.; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L589
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L589
framing: Test-fixture placeholder URL (host "x"); nothing to verify.
pass3 · url · high confidence
#28
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:884
"**Full evidence:** [trail](https://x).", "",
framing: Regex extractor matched a fake link inside a test string literal; nothing to check.; evidence: The "URL" `https://x` is a placeholder string inside a test fixture in a Python unit-test file (test_post_style_suggestions.py). It is dummy test data, not a published citation or a factual assertion.; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L884
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L884
framing: Regex extractor matched a fake link inside a test string literal; nothing to check.
pass3 · url · high confidence
#29
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:976
line = ("- Hyperscaler capex: ~$700B guided for 2026 across the major clouds. "
framing: Test-fixture string in tooling code, not a doc claim; the ~$700B hyperscaler capex figure was not checked because the line is sample input.; evidence: The text is a string literal (`line = (...)`) inside a Python unit-test file for the docs-review style-suggestion script. It is test input used to exercise the regex/suggestion logic. It is not published documentation or a factual…; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L976
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L976
framing: Test-fixture string in tooling code, not a doc claim; the ~$700B hyperscaler capex figure was not checked because the line is sample input.
pass3 · numerical · high confidence
#30
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:977
"Roughly 75% of it goes to AI. Agentic coding share: over 80% of code.")
framing: Sample text inside the test harness; nobody attributes it to an outside source.; evidence: The text comes from a Python test file (test_post_style_suggestions.py). The trailing `")` shows it is a string literal: sample prose used to test the style-suggestion script. It is not a published factual statement, so there is nothing…; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L977 (test fixture string)
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L977 (test fixture string)
framing: Sample text inside the test harness; nobody attributes it to an outside source.
pass3 · numerical · high confidence
#31
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:983
["Agentic coding share: roughly 90% of code by 2026."]),
evidence: This string is a test-fixture literal in a Python unit test for the docs-review style-suggestion script (test_post_style_suggestions.py). It is sample input for the tooling, not a published factual assertion, so the "90% by 2026" figure…; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L983 (test fixture list element); intuition: The figure looks like a deliberately unsupported sample stat that the style checker is meant to flag. It's only worth…
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L983 (test fixture list element)
pass3 · numerical · high confidence
#32
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:1007
"""Review of this mode, 2026-09-25: each case passed validate_entries and
framing: Internal test-code docstring; the date is a timestamp on the author's review, not a factual claim that needs a source.; evidence: This line is the opening of a docstring in the PR author's own test script. It says when the author reviewed their own test mode (2026-09-25) and that the cases passed validate_entries. It describes the author's own tooling, not a…; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L1007
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L1007
framing: Internal test-code docstring; the date is a timestamp on the author's review, not a factual claim that needs a source.
pass3 · numerical · high confidence
#33
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:1021
"""Review finding 2026-09-25: a #new-review renumbers from F1; keeping the
framing: Regex false positive: the date and 'F1' identifier in a test docstring matched the numerical pattern.; evidence: This line is a docstring in a Python test file (.claude/commands/docs-review/scripts/test_post_style_suggestions.py). It describes how the PR author's own review tooling numbers findings (a #new-review starts again from F1). The date…; source: repo:.claude/commands/docs-review/scripts/test_post_style_suggestions.py#L1021
source: repo:.claude/commands/docs-review/scripts/test_post_style_suggestions.py#L1021
framing: Regex false positive: the date and 'F1' identifier in a test docstring matched the numerical pattern.
pass3 · numerical · high confidence
#34
➖ Not a claim
.claude/commands/docs-review/scripts/test_post_style_suggestions.py:1038
"""Review finding 2026-09-25: lines matched on the checkout must not be
framing: The regex extractor matched a date inside a code-comment/docstring in a test file, so this is not published docs content.; evidence: The text is a Python test docstring in the PR's own internal review-tooling script. It records the date of an internal review finding ("Review finding 2026-09-25"). That date is metadata about the author's own pipeline, not a numerical…; source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L1038
source: .claude/commands/docs-review/scripts/test_post_style_suggestions.py:L1038
framing: The regex extractor matched a date inside a code-comment/docstring in a test file, so this is not published docs content.
pass3 · numerical · high confidence
#36
➖ Not a claim
.claude/commands/docs-review/scripts/validate-pinned.py:240
# 2026-09-25 the evidence line is an unprefixed column-0 `**Full evidence:**`,
framing: The regex matched the date stamp as a number. The comment records when the author's own output format changed, which is not a statistic.; evidence: This line is a code comment in the PR's own validation script. It describes the evidence-line format the author's review pipeline uses as of a dated change (2026-09-25). It's the author describing their own tooling, not a factual claim…; source: .claude/commands/docs-review/scripts/validate-pinned.py:L240
source: .claude/commands/docs-review/scripts/validate-pinned.py:L240
framing: The regex matched the date stamp as a number. The comment records when the author's own output format changed, which is not a statistic.
pass3 · numerical · high confidence
#37
➖ Not a claim
.claude/commands/docs-review/scripts/validate-pinned.py:2900
# `📎` is the pre-2026-09-25 prefix; live cards still carry it.
framing: The date is a changelog note about the author's own tooling, not a numerical claim about something outside it.; evidence: This is a code comment in the PR's own validation script. It describes the internal review pipeline's history, namely that the `📎` card prefix was used before 2026-09-25 and still shows up on live cards. It says nothing about a third…; source: repo:.claude/commands/docs-review/scripts/validate-pinned.py#L2900
source: repo:.claude/commands/docs-review/scripts/validate-pinned.py#L2900
framing: The date is a changelog note about the author's own tooling, not a numerical claim about something outside it.
pass3 · numerical · high confidence
#38
➖ Not a claim
.claude/commands/social-media-review/SKILL.md:56
After the per-platform reviews, if any platforms were repaired OR drafted, present the new copy under `### Suggested copy` with character counts…
evidence: This line is an instruction in the PR author's own review skill (.claude/commands/social-media-review/SKILL.md). It tells the skill how to format its output, using a collapsed <details> block for suggested copy. It makes no factual claim…; source: repo:.claude/commands/social-media-review/SKILL.md L56
source: repo:.claude/commands/social-media-review/SKILL.md L56
pass1 · temporal · high confidence
#39
➖ Not a claim
.claude/commands/social-media-review/SKILL.md:60
**In CI, open with one orienting line** directly under `## Social Media Review`, because authors who don't post to social media for work read this comment as…
framing: The regex picked up the date inside a design-rationale note. The reader feedback it refers to is internal and can't be checked, but the line describes the…; evidence: This line is an instruction inside an internal Claude skill file (SKILL.md). It describes how the author's own review comment should be formatted. The "numerical" token is just the date of internal reader feedback (2026-08-28), which is…; source: .claude/commands/social-media-review/SKILL.md L60 (the author's own pipeline design; no external source cited)
source: .claude/commands/social-media-review/SKILL.md L60 (the author's own pipeline design; no external source cited)
framing: The regex picked up the date inside a design-rationale note. The reader feedback it refers to is internal and can't be checked, but the line describes the author's own process rather than making a claim.
pass3 · numerical · medium confidence
#40
➖ Not a claim
.github/workflows/claude-code-review.yml:1626
# variable REVIEW_STYLE_INLINE is '1' (reader feedback, 2026-09-22:
evidence: This line is a YAML comment in a CI workflow. It describes the author's own config gate (REVIEW_STYLE_INLINE == '1') and gives the date of internal reader feedback. It isn't a factual assertion about something outside the repo. The regex…; source: .github/workflows/claude-code-review.yml:L1626 (code comment)
source: .github/workflows/claude-code-review.yml:L1626 (code comment)
pass3 · numerical · high confidence
#41
➖ Not a claim
.github/workflows/claude-code-review.yml:1731
# One-click fixes for what blocks merge (reader feedback, 2026-09-25:
framing: The regex flagged a date in a code comment that describes the PR author's own pipeline design.; evidence: This line is a YAML code comment in a GitHub workflow file. It records the author's own design note ("one-click fixes... reader feedback, 2026-09-25"). The date only says when the feedback came in. It isn't a numerical or factual claim…; source: .github/workflows/claude-code-review.yml:L1731
source: .github/workflows/claude-code-review.yml:L1731
framing: The regex flagged a date in a code comment that describes the PR author's own pipeline design.
pass3 · numerical · high confidence
#42
➖ Not a claim
.github/workflows/claude-code-review.yml:1893
local msg="$1" body
framing: The regex flagged the `1` in `$1` as a number, but it's a shell positional parameter.; evidence: This line is bash from a GitHub Actions workflow. It declares local variables, and `$1` is the function's first positional argument, not a number. There is nothing here to fact-check.; source: .github/workflows/claude-code-review.yml:L1893 (shell code line)
source: .github/workflows/claude-code-review.yml:L1893 (shell code line)
framing: The regex flagged the `1` in `$1` as a number, but it's a shell positional parameter.
pass3 · numerical · high confidence
#43
➖ Not a claim
.github/workflows/claude-social-review.yml:301
IMPORTANT: Keep the review short and scannable. Use `#### Platform — PASS` or `#### Platform — FAIL` headings with bullet-pointed reasons. No analysis…
framing: The regex flagged "30 seconds" as a number, but the line describes the author's own pipeline design and makes no claim about a third party.; evidence: This line is a prompt instruction in the PR author's own CI workflow (claude-social-review.yml). It tells the review bot how to format its output. "A blogger should read it in 30 seconds" is a design target for readability, not a…; source: .github/workflows/claude-social-review.yml:L301
source: .github/workflows/claude-social-review.yml:L301
framing: The regex flagged "30 seconds" as a number, but the line describes the author's own pipeline design and makes no claim about a third party.
pass3 · numerical · high confidence
#44
➖ Not a claim
.github/workflows/claude-update.yml:906
# One-click fixes for what blocks merge (reader feedback, 2026-09-25:
framing: The regex flagged the date in a comment as a numerical claim; it's only metadata describing the author's own workflow.; evidence: This line is a YAML code comment in the PR's own GitHub workflow. It notes the internal reader feedback that led to the design, and the date (2026-09-25) is just the author's timestamp. It isn't a falsifiable numerical or third-party claim.; source: .github/workflows/claude-update.yml:L906 (code comment, author's own pipeline design)
source: .github/workflows/claude-update.yml:L906 (code comment, author's own pipeline design)
framing: The regex flagged the date in a comment as a numerical claim; it's only metadata describing the author's own workflow.
pass3 · numerical · high confidence
#45
➖ Not a claim
.github/workflows/review-label-reconcile.yml:37
# uses, once the label is at least 30 minutes old (slack for API lag and
framing: The regex matched a number in a code comment. The value is a self-defined threshold, not an external claim.; evidence: This line is a YAML comment in the PR's own GitHub Actions workflow. It describes the author's chosen reconcile threshold: the label has to be at least 30 minutes old. It is a design parameter of the author's own pipeline, not a…; source: .github/workflows/review-label-reconcile.yml:L37 (code comment describing the author's own workflow design)
source: .github/workflows/review-label-reconcile.yml:L37 (code comment describing the author's own workflow design)
framing: The regex matched a number in a code comment. The value is a self-defined threshold, not an external claim.
pass3 · numerical · high confidence
#46
➖ Not a claim
CONTRIBUTING.md:63
While it works, your comment gets a 👀; when the card is updated, the 👀 becomes a 🚀. A failed refresh comments back to you with how to retry. `#new-review`…
evidence: This line is CONTRIBUTING.md describing how the repo's own review bot behaves (👀 while it works, 🚀 when done, a retry comment on failure, and the same signals for `#new-review`). That makes it the author describing their own pipeline…; source: gh pr diff 11 -R pulumi/docs-review-benchmarks (repo not resolvable); gh search code --owner pulumi "new-review rocket" (no results)
source: gh pr diff 11 -R pulumi/docs-review-benchmarks (repo not resolvable); gh search code --owner pulumi "new-review rocket" (no results)
pass1 · temporal · medium confidence
#47
➖ Not a claim
scripts/review-v3/README.md:224
the spinner, check-run and any `#new-review` confirmation (a requested run's
evidence: L224 is part of the README's own account of how the review-v3 pipeline handles a superseded handoff: "`redispatch` (head moved) ... cleans up the spinner, check-run and any `#new-review` confirmation". The regex matched `#new-review`…; source: repo:scripts/review-v3/README.md L219-227
source: repo:scripts/review-v3/README.md L219-227
pass1 · temporal · high confidence
#48
➖ Not a claim
scripts/review-v3/collect.py:267
# 💡/📎 are the pre-2026-09-25 prefixes of the two lines after the list.
evidence: This is a code comment in the PR author's own review script (scripts/review-v3/collect.py). It describes the emoji prefixes (💡/📎) that the pipeline's own output used before 2026-09-25. That's the author describing their own design and…; source: scripts/review-v3/collect.py:L267 (code comment describing internal pipeline format)
source: scripts/review-v3/collect.py:L267 (code comment describing internal pipeline format)
pass3 · numerical · high confidence
No claims match the current filter.
Findings
No findings.
Editorial stances
Superlative, ranking, or comparative language the diff adds. No verdict — a page's own framing isn't fact-checkable — listed so a reviewer can confirm each is a stance the docs should take.
- .claude/commands/docs-review/references/update.md L344 - `summary` is the only way to change the card's italic one-sentence summary. positioning · found by regex
Investigation log
- cross-sibling-reads
- not run (not in a templated section)
- external-claim-verification
- 1 of 48 claims verified (0 unverifiable, 0 contradicted) · 4 specialists (numerical, cross-reference, capability, framing); 0 cross-specialist corroborations · routed: 0 inline, 7 Pass 1, 0 Pass 2, 41 Pass 3 (verified 0, contradicted 0, unverifiable 41).
- cited-claim-spot-checks
- not run (no cited claims)
- frontmatter-sweep
- not run (no frontmatter in diff)
- temporal-trigger-sweep
- ran (recency words present in diff; spot-check in-review)
- code-execution
- not run (no `static/programs/` change)
- code-examples-checks
- ran (3 specialists: structural, existence, body-code-coverage); 0 findings
- editorial-balance-pass
- not run (not under content/blog/)
History
- 2026-09-25T23:16:07Z
62384bf
initial review (pending publication)