update examples by IceyLiu · Pull Request #3 · iOfficeAI/OfficeCLI

IceyLiu · 2026-03-21T21:53:26Z

updated to a new batch of styles
deleted the redundant documents

… subprocess Python can call OfficeCli via subprocess with JSON output parsed by json.loads(), making integration cost much lower than initially assessed. Updated rankings: OfficeCli moves from #4 to #3 in traditional Python development scenario. https://claude.ai/code/session_012F47kVKAP63YQe6hdWhc1b

update examples

… subprocess Python can call OfficeCli via subprocess with JSON output parsed by json.loads(), making integration cost much lower than initially assessed. Updated rankings: OfficeCli moves from #4 to #3 in traditional Python development scenario. https://claude.ai/code/session_012F47kVKAP63YQe6hdWhc1b

update examples

…dHeaders) Before: the header was emitted only on page 1 and always used whichever HeaderPart happened to be first in the document. The footer likewise ignored the first/even/default variants. This breaks virtually every formal Chinese technical report (GB/T style), theses, and legal docs, all of which rely on: - <w:titlePg/> on the section → first-page header is distinct (封面). - <w:evenAndOddHeaders/> in document settings → even/odd pages swap. Changes: - BuildSectionHfBundles resolves each section's HeaderReference / FooterReference elements, buckets them by type (first / default / even), and pre-renders each to an HTML string keyed by section idx. - PickHeaderFooter applies the selection rules per page: * first-page-of-section + titlePg → First * evenAndOddHeaders on + even page → Even * otherwise → Default (legacy fallback when the section defines no variants). - Page-emit loop now tracks activeSectionIdx (via existing SECT markers) and whether the current page is the first of that section. Header now lands on every page, not just page 1. - Footer PAGE / NUMPAGES substitution runs against the picked footer variant, not the global one, so even-footer page numbers render too. Fixes KNOWN_ISSUES.md #3.

…er bugs - tester #2 (JS): tblHeader continuation iterated table.querySelectorAll('tr') which pulled nested subtable rows into the split, mangling nested structures. Filter to top-level rows via tr.closest('table')===table. - tester #3 (JS): restart='newSection' compared only prev/current restart value, so two adjacent sections both declaring newSection never triggered reset. Page-wrapper now carries data-section-idx; JS resets the running counter on section boundary, not on restart-value change. - tester #4: altChunk ContentType equality failed on 'text/html; charset=utf-8' and other '+xml' variants — HTML payload fell through to the RTF stripper. Split media-type from params and accept xhtml+xml variants. - fuzzer A: pgSz w/h parsed via UInt32Value.Value crashed on negative raw attrs. Wrap Width/Height access in a swallow-to-fallback helper. - fuzzer B/C: lnNumType CountBy/Start/Distance typed as Int16Value; malformed raw values (non-numeric, overflow) threw on .Value. Parse InnerText manually via short.TryParse / int.TryParse. - fuzzer D/E: huge w:ilvl (10000, Int32.MaxValue) either popped an empty open-tag stack (crash) or inflated the HTML by 50× per paragraph (DoS). Clamp ilvl to OOXML-legal [0,8].

…light, drop static contrast Gate 4 Min cycle and Delivery Gate were two passes at the same checks: Min cycle #2 (view html) overlapped Gate 5, #3 (token query) overlapped Gate 2's grep, #4 (close+validate) overlapped Gate 1, #5 (target viewer) was a 4-line restatement of the view-html-vs-runtime rule already in Common Workflow #6. The intro itself called Min cycle "warmup" while Delivery Gate did the real work. Replace the 6-step Min cycle with a single-line Pre-flight using view annotated (the one Min-cycle item Gates don't cover — font/size violations and overflow surfaced before the gate run). Move target viewer + fix-and-rerun guidance into a tight "After all gates pass" subsection at the end of QA. Drop the static-contrast Gate 4 (query+jq dark-fill scan): its hardcoded list of five hex values misses any deck using a custom dark fill, while Gate 5's visual review catches dark-on-dark across all fills. Renumber: Gate 5 → Gate 4. Update cross-references (intro line, status banner, "Gates 1-4 cannot see rendered slides" → "Gates 1-3", "Gates 2-5 are how you catch" → "Gates 2-4").

Gate 1 was a 4-line wrapper around officecli validate that captured output, ran a redundant grep on it, and re-printed it on failure — validate already exits non-zero on schema error and prints its own output. Reduce to a single line with the canonical pattern. Drop Gate 2 (token-leak grep) entirely. Most of its branches ($VAR$, {{name}}, lorem, xxxx, <TODO>) catch failure modes from template-substitution workflows that this skill does not teach; agents that build from real user content never produce them. The remaining branches (\$\t\n shell-escape literals and empty parens/brackets from shell-stripped chart titles) are visible in Gate 3's per-slide HTML walk and are also prevented at write time by the Shell & Execution Discipline section. Renumber: Gate 3 (slide order) → Gate 2, Gate 4 (visual audit) → Gate 3. Update all cross-references — Common Workflow #3, recipe (d) reference, intro line ("Three checks"), banner messages, and the closing line. Drop the Common Pitfalls row about "[ ] empty-bracket checkboxes false-positive in Gate 2" since the check it warned about is gone.

@name

The previous Gate 2 dumped slide titles via query+jq and asked the agent to manually compare against the build plan — same subjective mental check that Gate 3's order-sanity item already covers. Worse, the first query (shape[@name=Title]) silently fell through every time because no recipe in this skill names title shapes "Title", so the gate landed on a fallback that printed all shapes truncated at 40 entries. Replace with a deterministic view-issues pass: run officecli view issues, filter the expected layout=blank "Slide has no title" noise, REJECT on any remaining [X#] issue. This catches real problems (text overflow, structure issues) automatically instead of asking the agent to spot them. Slide-order audit moves entirely to Gate 3's order-sanity question. Update Common Workflow #3 and recipe (d') cross-references to point at Gate 3 instead of Gate 2. Update the intro line and the closing "Gates 2-3 are how you catch X, Y, Z" sentence.

update examples

…dHeaders) Before: the header was emitted only on page 1 and always used whichever HeaderPart happened to be first in the document. The footer likewise ignored the first/even/default variants. This breaks virtually every formal Chinese technical report (GB/T style), theses, and legal docs, all of which rely on: - <w:titlePg/> on the section → first-page header is distinct (封面). - <w:evenAndOddHeaders/> in document settings → even/odd pages swap. Changes: - BuildSectionHfBundles resolves each section's HeaderReference / FooterReference elements, buckets them by type (first / default / even), and pre-renders each to an HTML string keyed by section idx. - PickHeaderFooter applies the selection rules per page: * first-page-of-section + titlePg → First * evenAndOddHeaders on + even page → Even * otherwise → Default (legacy fallback when the section defines no variants). - Page-emit loop now tracks activeSectionIdx (via existing SECT markers) and whether the current page is the first of that section. Header now lands on every page, not just page 1. - Footer PAGE / NUMPAGES substitution runs against the picked footer variant, not the global one, so even-footer page numbers render too. Fixes KNOWN_ISSUES.md iOfficeAI#3.

…er bugs - tester iOfficeAI#2 (JS): tblHeader continuation iterated table.querySelectorAll('tr') which pulled nested subtable rows into the split, mangling nested structures. Filter to top-level rows via tr.closest('table')===table. - tester iOfficeAI#3 (JS): restart='newSection' compared only prev/current restart value, so two adjacent sections both declaring newSection never triggered reset. Page-wrapper now carries data-section-idx; JS resets the running counter on section boundary, not on restart-value change. - tester iOfficeAI#4: altChunk ContentType equality failed on 'text/html; charset=utf-8' and other '+xml' variants — HTML payload fell through to the RTF stripper. Split media-type from params and accept xhtml+xml variants. - fuzzer A: pgSz w/h parsed via UInt32Value.Value crashed on negative raw attrs. Wrap Width/Height access in a swallow-to-fallback helper. - fuzzer B/C: lnNumType CountBy/Start/Distance typed as Int16Value; malformed raw values (non-numeric, overflow) threw on .Value. Parse InnerText manually via short.TryParse / int.TryParse. - fuzzer D/E: huge w:ilvl (10000, Int32.MaxValue) either popped an empty open-tag stack (crash) or inflated the HTML by 50× per paragraph (DoS). Clamp ilvl to OOXML-legal [0,8].

…light, drop static contrast Gate 4 Min cycle and Delivery Gate were two passes at the same checks: Min cycle iOfficeAI#2 (view html) overlapped Gate 5, iOfficeAI#3 (token query) overlapped Gate 2's grep, iOfficeAI#4 (close+validate) overlapped Gate 1, iOfficeAI#5 (target viewer) was a 4-line restatement of the view-html-vs-runtime rule already in Common Workflow iOfficeAI#6. The intro itself called Min cycle "warmup" while Delivery Gate did the real work. Replace the 6-step Min cycle with a single-line Pre-flight using view annotated (the one Min-cycle item Gates don't cover — font/size violations and overflow surfaced before the gate run). Move target viewer + fix-and-rerun guidance into a tight "After all gates pass" subsection at the end of QA. Drop the static-contrast Gate 4 (query+jq dark-fill scan): its hardcoded list of five hex values misses any deck using a custom dark fill, while Gate 5's visual review catches dark-on-dark across all fills. Renumber: Gate 5 → Gate 4. Update cross-references (intro line, status banner, "Gates 1-4 cannot see rendered slides" → "Gates 1-3", "Gates 2-5 are how you catch" → "Gates 2-4").

Gate 1 was a 4-line wrapper around officecli validate that captured output, ran a redundant grep on it, and re-printed it on failure — validate already exits non-zero on schema error and prints its own output. Reduce to a single line with the canonical pattern. Drop Gate 2 (token-leak grep) entirely. Most of its branches ($VAR$, {{name}}, lorem, xxxx, <TODO>) catch failure modes from template-substitution workflows that this skill does not teach; agents that build from real user content never produce them. The remaining branches (\$\t\n shell-escape literals and empty parens/brackets from shell-stripped chart titles) are visible in Gate 3's per-slide HTML walk and are also prevented at write time by the Shell & Execution Discipline section. Renumber: Gate 3 (slide order) → Gate 2, Gate 4 (visual audit) → Gate 3. Update all cross-references — Common Workflow iOfficeAI#3, recipe (d) reference, intro line ("Three checks"), banner messages, and the closing line. Drop the Common Pitfalls row about "[ ] empty-bracket checkboxes false-positive in Gate 2" since the check it warned about is gone.

@name

The previous Gate 2 dumped slide titles via query+jq and asked the agent to manually compare against the build plan — same subjective mental check that Gate 3's order-sanity item already covers. Worse, the first query (shape[@name=Title]) silently fell through every time because no recipe in this skill names title shapes "Title", so the gate landed on a fallback that printed all shapes truncated at 40 entries. Replace with a deterministic view-issues pass: run officecli view issues, filter the expected layout=blank "Slide has no title" noise, REJECT on any remaining [X#] issue. This catches real problems (text overflow, structure issues) automatically instead of asking the agent to spot them. Slide-order audit moves entirely to Gate 3's order-sanity question. Update Common Workflow iOfficeAI#3 and recipe (d') cross-references to point at Gate 3 instead of Gate 2. Update the intro line and the closing "Gates 2-3 are how you catch X, Y, Z" sentence.

…dHeaders) Before: the header was emitted only on page 1 and always used whichever HeaderPart happened to be first in the document. The footer likewise ignored the first/even/default variants. This breaks virtually every formal Chinese technical report (GB/T style), theses, and legal docs, all of which rely on: - <w:titlePg/> on the section → first-page header is distinct (封面). - <w:evenAndOddHeaders/> in document settings → even/odd pages swap. Changes: - BuildSectionHfBundles resolves each section's HeaderReference / FooterReference elements, buckets them by type (first / default / even), and pre-renders each to an HTML string keyed by section idx. - PickHeaderFooter applies the selection rules per page: * first-page-of-section + titlePg → First * evenAndOddHeaders on + even page → Even * otherwise → Default (legacy fallback when the section defines no variants). - Page-emit loop now tracks activeSectionIdx (via existing SECT markers) and whether the current page is the first of that section. Header now lands on every page, not just page 1. - Footer PAGE / NUMPAGES substitution runs against the picked footer variant, not the global one, so even-footer page numbers render too. Fixes KNOWN_ISSUES.md #3.

…er bugs - tester #2 (JS): tblHeader continuation iterated table.querySelectorAll('tr') which pulled nested subtable rows into the split, mangling nested structures. Filter to top-level rows via tr.closest('table')===table. - tester #3 (JS): restart='newSection' compared only prev/current restart value, so two adjacent sections both declaring newSection never triggered reset. Page-wrapper now carries data-section-idx; JS resets the running counter on section boundary, not on restart-value change. - tester #4: altChunk ContentType equality failed on 'text/html; charset=utf-8' and other '+xml' variants — HTML payload fell through to the RTF stripper. Split media-type from params and accept xhtml+xml variants. - fuzzer A: pgSz w/h parsed via UInt32Value.Value crashed on negative raw attrs. Wrap Width/Height access in a swallow-to-fallback helper. - fuzzer B/C: lnNumType CountBy/Start/Distance typed as Int16Value; malformed raw values (non-numeric, overflow) threw on .Value. Parse InnerText manually via short.TryParse / int.TryParse. - fuzzer D/E: huge w:ilvl (10000, Int32.MaxValue) either popped an empty open-tag stack (crash) or inflated the HTML by 50× per paragraph (DoS). Clamp ilvl to OOXML-legal [0,8].

…light, drop static contrast Gate 4 Min cycle and Delivery Gate were two passes at the same checks: Min cycle #2 (view html) overlapped Gate 5, #3 (token query) overlapped Gate 2's grep, #4 (close+validate) overlapped Gate 1, #5 (target viewer) was a 4-line restatement of the view-html-vs-runtime rule already in Common Workflow #6. The intro itself called Min cycle "warmup" while Delivery Gate did the real work. Replace the 6-step Min cycle with a single-line Pre-flight using view annotated (the one Min-cycle item Gates don't cover — font/size violations and overflow surfaced before the gate run). Move target viewer + fix-and-rerun guidance into a tight "After all gates pass" subsection at the end of QA. Drop the static-contrast Gate 4 (query+jq dark-fill scan): its hardcoded list of five hex values misses any deck using a custom dark fill, while Gate 5's visual review catches dark-on-dark across all fills. Renumber: Gate 5 → Gate 4. Update cross-references (intro line, status banner, "Gates 1-4 cannot see rendered slides" → "Gates 1-3", "Gates 2-5 are how you catch" → "Gates 2-4").

Gate 1 was a 4-line wrapper around officecli validate that captured output, ran a redundant grep on it, and re-printed it on failure — validate already exits non-zero on schema error and prints its own output. Reduce to a single line with the canonical pattern. Drop Gate 2 (token-leak grep) entirely. Most of its branches ($VAR$, {{name}}, lorem, xxxx, <TODO>) catch failure modes from template-substitution workflows that this skill does not teach; agents that build from real user content never produce them. The remaining branches (\$\t\n shell-escape literals and empty parens/brackets from shell-stripped chart titles) are visible in Gate 3's per-slide HTML walk and are also prevented at write time by the Shell & Execution Discipline section. Renumber: Gate 3 (slide order) → Gate 2, Gate 4 (visual audit) → Gate 3. Update all cross-references — Common Workflow #3, recipe (d) reference, intro line ("Three checks"), banner messages, and the closing line. Drop the Common Pitfalls row about "[ ] empty-bracket checkboxes false-positive in Gate 2" since the check it warned about is gone.

@name

The previous Gate 2 dumped slide titles via query+jq and asked the agent to manually compare against the build plan — same subjective mental check that Gate 3's order-sanity item already covers. Worse, the first query (shape[@name=Title]) silently fell through every time because no recipe in this skill names title shapes "Title", so the gate landed on a fallback that printed all shapes truncated at 40 entries. Replace with a deterministic view-issues pass: run officecli view issues, filter the expected layout=blank "Slide has no title" noise, REJECT on any remaining [X#] issue. This catches real problems (text overflow, structure issues) automatically instead of asking the agent to spot them. Slide-order audit moves entirely to Gate 3's order-sanity question. Update Common Workflow #3 and recipe (d') cross-references to point at Gate 3 instead of Gate 2. Update the intro line and the closing "Gates 2-3 are how you catch X, Y, Z" sentence.

update

60bbbb2

IceyLiu merged commit f52f4bc into main Mar 21, 2026

goworm pushed a commit that referenced this pull request Mar 27, 2026

Merge pull request #3 from iOfficeAI/docs

86271a6

update examples

goworm pushed a commit that referenced this pull request Apr 2, 2026

Merge pull request #3 from iOfficeAI/docs

434ebd9

update examples

goworm mentioned this pull request May 11, 2026

发现可能是不完美的地方 #104

Closed

NextDoorLaoHuang-HF pushed a commit to NextDoorLaoHuang-HF/OfficeCLI that referenced this pull request Jun 10, 2026

Merge pull request iOfficeAI#3 from iOfficeAI/docs

d2f4d1d

update examples

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

update examples#3

update examples#3
IceyLiu merged 1 commit into
mainfrom
docs

IceyLiu commented Mar 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

IceyLiu commented Mar 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant