New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Spreadsheet Proof Gate

by HKUDS in HKUDS/OpenSpace

Edit Excel workbooks only after anchored preflight audit, then require deterministic Python/openpyxl post-write proof that every requested criterion is satisfied or explicitly reconciled before…

MITAuto-check passedDocuments & Office

Install Spreadsheet Proof Gate

skills CLI
$ npx skills add HKUDS/OpenSpace --skill spreadsheet-proof-gate -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install HKUDS/OpenSpace spreadsheet-proof-gate --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/HKUDS/OpenSpace.git skills-src && mkdir -p .claude/skills && cp -r skills-src/benchmarks/gdpval/skills/spreadsheet-audit-edit-validate-merged .claude/skills/spreadsheet-proof-gate && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
spreadsheet-proof-gate
GitHub stars
7.7k
Token cost
~4.5k tokens
SKILL.md length
2,087 words
Files
2
Skills in repo
199
Repo updated
First seen
Licence
MIT

At a glance

Edit Excel workbooks only after anchored preflight audit, then require deterministic Python/openpyxl post-write proof that every requested criterion is satisfied or explicitly reconciled before…

  • Works in 10 steps: Discover candidate workbooks → Convert the user request into explicit… → Pre-edit audit → …
  • Tasks that involve Excel spreadsheets
  • SKILL.md covers When to use, Core rule, Workspace anchoring rule and Outcome contract, plus 16 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Spreadsheet Proof Gate is an agent skill from HKUDS/OpenSpace. Edit Excel workbooks only after anchored preflight audit, then require deterministic Python/openpyxl post-write proof that every requested criterion is satisfied or explicitly reconciled before finalizing.

Its SKILL.md is about 4.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 1 other file.

It sits in Documents & Office, covering Excel spreadsheets. It works with Microsoft Excel, Python and openpyxl. The repository describes itself as: "OpenSpace: The Skill Management Layer for AI Agents" -- https://open-space.cloud/. The licence is MIT.

When your agent uses it

  • Tasks that involve Excel spreadsheets

Example prompts

  • “/spreadsheet-proof-gate”

Requirements

  • Python 3

Workflow steps

10 steps, taken from the step headings in SKILL.md.

  1. Discover candidate workbooks
  2. Convert the user request into explicit criteria
  3. Pre-edit audit
  4. Pre-edit go/no-go gate
  5. Perform the edit
  6. Deterministic post-write proof
  7. Build the criteria list
  8. Run a direct Python inspection
  9. Add task-specific criteria to the script
  10. Reconcile non-pass results

What it can do on your machine

Read from SKILL.md and the folder at commit 3827781. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are python).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Spreadsheet Proof Gate loads about 4.5k tokens when it runs. Until then it costs about 57 tokens; SKILL.md has 2,087 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~57
When it runs · the whole SKILL.md, loaded when a task matches
~4.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from HKUDS/OpenSpace at commit 3827781, republished under its MIT licence (© HKUDS). 2,087 words, ~4,475 tokens.

Download SKILL.mdSave it as .claude/skills/spreadsheet-proof-gate/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
spreadsheet-proof-gate
description
Edit Excel workbooks only after anchored preflight audit, then require deterministic Python/openpyxl post-write proof that every requested criterion is satisfied or explicitly reconciled before finalizing.

Spreadsheet Proof Gate

Use this skill when you must modify, create, or validate an Excel workbook and the result must be proven against a user specification rather than merely described.

This skill combines:

  • workbook discovery and identity confirmation
  • pre-edit structure auditing
  • safe workbook editing
  • deterministic post-write verification from disk
  • a machine-checkable pass/fail checklist for every requested criterion
  • a hard finalization gate that rejects completion when any required criterion is unmet or unreconciled

This is a workflow skill. It does not require a specific editor, but it does require a direct Python verification pass with openpyxl before you report success.

When to use

Use this skill when any of the following are true:

  • the workbook path must be discovered
  • multiple similar Excel files may exist
  • the user requires exact sheet names, columns, counts, or samples
  • the task edits an existing workbook
  • a delegated agent may perform spreadsheet work
  • prior summaries are inconsistent or not trustworthy
  • exact post-edit workbook state matters
  • you must prove that mandatory criteria were satisfied

Core rule

Never finalize from narrative summary alone.

A workbook task is only complete when:

  1. the target workbook was identified from the anchored workspace path
  2. pre-edit structure matched the intended operation
  3. the edit was performed on the correct file
  4. the saved workbook was directly re-read from disk
  5. every requested criterion was checked in a deterministic verification report
  6. each criterion is marked as one of:
    • PASS
    • FAIL
    • UNAVAILABLE-IN-SOURCE
  7. any non-pass result is explicitly reconciled before finalizing

If any mandatory criterion is still FAIL, do not declare success.

If a criterion is UNAVAILABLE-IN-SOURCE, only finalize if:

  • the source workbook truly lacks the needed input,
  • you state that clearly,
  • and the user request did not require inventing unsupported data.

Workspace anchoring rule

All discovery, reads, writes, and verification must be anchored to the exact workspace path provided for the task.

Rules:

  • treat the provided workspace path as the authoritative root
  • resolve the workbook to an exact path under that root
  • do not silently switch to a similarly named file elsewhere
  • do not trust a delegated tool's reported path without independently confirming it
  • if the anchored file cannot be found, stop rather than guessing

Outcome contract

Your output to yourself and to the user should be based on proof, not inference.

For every task, maintain three artifacts internally:

  1. Candidate inventory
  2. Pre-edit audit checklist
  3. Post-write proof checklist

The post-write proof checklist is decisive.

Workflow

Phase 1: Discover candidate workbooks

Search under the exact workspace root for plausible Excel files:

  • .xlsx
  • .xlsm
  • .xls

Prefer files that:

  • match names mentioned by the user
  • live in likely data/output/project folders
  • have relevant modification times
  • contain expected sheet names
  • contain plausible row/column structure

If several files are plausible, inspect all plausible candidates before choosing.

Minimum evidence before selecting a workbook

Before choosing the workbook to edit, confirm at least:

  • exact anchored path
  • file existence
  • workbook readability
  • sheet names
  • approximate row counts in relevant sheets
  • relevant headers in target sheets

Do not select a workbook based on filename similarity alone.

Phase 2: Convert the user request into explicit criteria

Before editing, extract every requirement you can verify.

Possible criteria include:

  • required workbook path or output path
  • required sheet names
  • sheets to preserve
  • sheets to create
  • target sheet to edit
  • required columns
  • required formulas or summaries
  • minimum data rows
  • exact row counts
  • sample size
  • marker/flag count
  • selection rules
  • category coverage requirements
  • preservation of macros/formatting
  • preservation of untouched sheets

Turn these into an explicit checklist.

Criteria format

Represent each requested requirement as a discrete checkable item.

Good examples:

  • required-sheet: Summary exists
  • required-sheet: Sample exists
  • required-column: Sheet Sample has column "Selected"
  • sample-count: exactly 25 marked rows in Sample
  • coverage: at least 1 marked row where Product = Trade Finance
  • preservation: source sheet RawData still exists
  • formula: column H contains formulas for all populated rows

Bad examples:

  • workbook looks right
  • sampling seems okay
  • most tabs present

Phase 3: Pre-edit audit

Before any write, inspect the workbook structure against the extracted criteria.

Required pre-edit checks

For each relevant workbook:

  1. confirm exact anchored path
  2. list all sheet names
  3. identify target sheets
  4. inspect headers
  5. count relevant rows
  6. note formulas, merged cells, tables, filters, protections, or macros if relevant
  7. confirm whether required columns exist already or must be created
  8. confirm whether the planned edit is structurally safe
Identity consistency check

If you inspect the workbook multiple times or with multiple tools, the inspections must agree on:

  • exact path
  • sheet names
  • relevant row counts
  • used range or dimensions when available
  • relevant headers

If they disagree, treat workbook identity as unconfirmed.

Conflict reconciliation rule

If two inspections disagree on sheet names, row counts, dimensions, or headers, do not edit until you can produce all of the following:

  • an independent direct read from the exact anchored path
  • explicit evidence that this is the intended workbook
  • a reconciled explanation of which earlier inspection was wrong and why
  • a fresh verified inventory of sheets, dimensions, row counts, and headers from the anchored file

If you cannot reconcile the conflict, the decision is no-go.

Phase 4: Pre-edit go/no-go gate

Proceed only if all are true:

  • the workbook identity is confirmed
  • the target sheet is unambiguous
  • required columns exist or can be added safely
  • row counts are plausible for the requested operation
  • no unresolved inspection conflicts remain
  • the requested edit is possible from the available source data

Otherwise stop and report the mismatch.

Phase 5: Perform the edit

Only after the workbook passes pre-edit audit:

  • edit only the intended workbook
  • preserve untouched sheets unless instructed otherwise
  • preserve names unless renaming was requested
  • preserve formatting/macros/formulas when required
  • keep a clear mapping from user requirements to written cells/rows

Record at minimum:

  • workbook path edited
  • output path written
  • sheets modified
  • columns added or populated
  • rows added or updated
  • formulas inserted
  • rows marked/selected
  • any assumptions made

Phase 6: Deterministic post-write proof

This phase is mandatory.

After saving, directly inspect the saved workbook from disk with Python and openpyxl. Do not finalize from memory, delegated prose, or a generic success message.

The verification pass must produce a compact, machine-checkable checklist covering every extracted criterion.

Verification standard

The verification must be:

  • direct: reads the workbook from disk
  • deterministic: script-based, not conversational
  • anchored: uses exact workspace paths
  • criterion-based: checks each requirement explicitly
  • decisive: any unresolved failure blocks completion

Required proof outputs

At minimum, the verification must report:

  • inspected file path
  • whether the file exists
  • workbook sheet names
  • per-sheet headers for relevant sheets
  • per-sheet non-empty row counts
  • marker/selected-row counts if relevant
  • formula presence if relevant
  • preservation of required source sheets
  • existence of requested output files
  • criterion checklist with status for each item

Status vocabulary

Each criterion must end in exactly one status:

  • PASS — requirement satisfied by direct inspection
  • FAIL — requirement not satisfied
  • UNAVAILABLE-IN-SOURCE — requirement could not be satisfied because required source information was absent from the workbook or inputs

Do not replace these with softer wording like:

  • looks okay
  • appears complete
  • probably satisfied
  • seems unavailable

Finalization gate

You may finalize only when every required criterion is either:

  • PASS, or
  • UNAVAILABLE-IN-SOURCE with explicit reconciliation

You must not finalize when:

  • any mandatory criterion remains FAIL
  • a claimed output file was not verified from disk
  • the workbook path is uncertain
  • post-write proof was not run
  • you only have delegated summary evidence
  • requested sample counts and actual marked counts differ without reconciliation
  • required coverage criteria are missing and not repaired
  • required sheet names or headers are missing

If direct proof says the workbook is incomplete, that proof is authoritative.

Show full SKILL.md (839 more words)Show less

Step 1: Build the criteria list

Create a numbered list of all checkable requirements from the user request.

Example:

  1. output workbook exists at <path>
  2. sheet Sample exists
  3. sheet Summary exists
  4. Sample has column Selected
  5. exactly 25 rows are marked selected
  6. at least one selected row has Product = Trade Finance
  7. source sheet RawData still exists

Step 2: Run a direct Python inspection

Use a standalone script or inline Python. openpyxl is the default.

Verification script requirements

Your script should print, in a compact format:

  • file path
  • existence
  • sheet list
  • relevant headers
  • relevant row counts
  • relevant marked-row counts
  • any criterion-specific counts
  • one line per criterion with a status

Adapt this template to the task.

python
from pathlib import Path
from openpyxl import load_workbook

WORKBOOK = Path("TARGET.xlsx")

TRUTHY = {"x", "yes", "true", "1", "y"}
MARKER_CANDIDATES = {"mark", "marked", "flag", "selected", "include", "status"}

def norm(v):
    if v is None:
        return ""
    return str(v).strip()

def lower(v):
    return norm(v).lower()

def first_nonempty_row(ws, max_scan=20):
    for r in ws.iter_rows(min_row=1, max_row=min(ws.max_row, max_scan), values_only=True):
        vals = list(r)
        if any(norm(v) != "" for v in vals):
            return vals
    return []

def data_rows(ws, header_row_idx=1):
    count = 0
    for row in ws.iter_rows(min_row=header_row_idx + 1, values_only=True):
        if any(norm(v) != "" for v in row):
            count += 1
    return count

def marker_index(headers):
    normalized = [lower(h) for h in headers]
    for i, h in enumerate(normalized):
        if h in MARKER_CANDIDATES:
            return i
    return None

def is_truthy(v):
    return lower(v) in TRUTHY

def marked_rows(ws, header_row_idx=1, idx=None):
    if idx is None:
        return 0
    count = 0
    for row in ws.iter_rows(min_row=header_row_idx + 1, values_only=True):
        if not any(norm(v) != "" for v in row):
            continue
        value = row[idx] if idx < len(row) else None
        if is_truthy(value):
            count += 1
    return count

criteria = []

print(f"WORKBOOK: {WORKBOOK}")
print(f"EXISTS: {WORKBOOK.exists()}")

if not WORKBOOK.exists():
    print("CRITERION|output-exists|FAIL|Workbook file missing")
    raise SystemExit(0)

wb = load_workbook(WORKBOOK, data_only=False)
print(f"SHEETS: {wb.sheetnames}")

for ws in wb.worksheets:
    headers = first_nonempty_row(ws)
    idx = marker_index(headers) if headers else None
    print(f"SHEET|{ws.title}|HEADERS|{headers}")
    print(f"SHEET|{ws.title}|DATA_ROWS|{data_rows(ws, header_row_idx=1)}")
    if idx is not None:
        print(f"SHEET|{ws.title}|MARKER_COLUMN|{headers[idx]}")
        print(f"SHEET|{ws.title}|MARKED_ROWS|{marked_rows(ws, header_row_idx=1, idx=idx)}")
    else:
        print(f"SHEET|{ws.title}|MARKER_COLUMN|<not found>")

# Add explicit criteria below. Examples:
required_sheets = ["Sample", "Summary"]
for s in required_sheets:
    status = "PASS" if s in wb.sheetnames else "FAIL"
    reason = "present" if status == "PASS" else "missing"
    print(f"CRITERION|required-sheet:{s}|{status}|{reason}")

Step 3: Add task-specific criteria to the script

The generic script is not enough by itself. Extend it to test the actual request, such as:

  • exact sample count
  • marker count by sheet
  • presence of category coverage
  • formulas in a target column
  • exact column names
  • output file existence
  • preservation of source sheets
  • expected row count delta
  • values meeting business rules

If the task depends on category coverage, test category coverage explicitly rather than assuming it from total counts.

Example criteria additions:

  • exactly 40 marked rows
  • at least one marked row for each required region
  • column "Review Note" populated for all marked rows
  • sheet "Summary" cell B2 contains numeric total
  • sheet "RawData" still present after save

Step 4: Reconcile non-pass results

For each non-pass criterion:

If status is FAIL

Do one of:

  • repair the workbook and re-run proof
  • report failure clearly
  • ask for clarification if the request was ambiguous

Do not declare completion while a mandatory FAIL remains.

If status is UNAVAILABLE-IN-SOURCE

State explicitly:

  • what source input was missing
  • which criterion could not be satisfied from source data
  • whether the task allowed this limitation
  • whether the workbook was left unchanged or partially completed because of it

Use UNAVAILABLE-IN-SOURCE only for true source limitations, not for editing mistakes.

Required machine-checkable checklist

Every post-write verification must include a checklist in this style:

  • CRITERION|<name>|PASS|<reason>
  • CRITERION|<name>|FAIL|<reason>
  • CRITERION|<name>|UNAVAILABLE-IN-SOURCE|<reason>

This format is intentionally rigid so you can reason from concrete output.

Checklist design rules

  • one criterion per line
  • criterion names should be specific
  • reasons should be short and factual
  • do not collapse multiple requirements into one broad criterion
  • if a requirement is critical, give it its own line

Decision rules

Hard no-go conditions before editing

Do not edit if:

  • workbook path is unconfirmed
  • candidate workbooks remain ambiguous
  • required target sheet cannot be identified
  • repeated inspections conflict and are unreconciled
  • the workbook cannot be read from the anchored path

Hard fail conditions after editing

Do not finalize if:

  • a required criterion is FAIL
  • sample count differs from request and is unexplained
  • required category or rule-based coverage is missing
  • direct verification cannot read the saved workbook
  • only delegated-agent summary evidence exists
  • required output path was not verified
  • required source sheet disappeared unexpectedly

When delegated summaries disagree with direct proof

Trust direct proof from the deterministic script.

If a delegated agent claimed success but your direct verification shows missing sheets, wrong counts, missing categories, or missing output files, the task is not complete.

Practical checklist

Before editing

  • anchor to the exact workspace root
  • locate plausible Excel files
  • inspect all plausible candidates
  • record exact target path
  • list sheet names
  • inspect relevant headers
  • count relevant rows
  • detect inspection conflicts
  • reconcile any conflict by direct anchored read
  • convert the user request into explicit criteria
  • decide go/no-go

After editing

  • save the workbook to the intended path
  • re-open it from disk
  • run deterministic Python/openpyxl verification
  • print workbook facts
  • print one criterion line per requirement
  • repair any FAIL if safe
  • re-run verification
  • finalize only if every mandatory item is PASS or explicitly reconciled UNAVAILABLE-IN-SOURCE

Final response pattern

When reporting to the user, base your summary only on verified facts from the direct proof output.

Include:

  • exact workbook path verified
  • whether the output file exists
  • sheet names present
  • relevant row counts
  • marked-row counts if relevant
  • the outcome of mandatory criteria
  • any remaining limitation explicitly

Preferred phrasing:

  • "I verified the saved workbook directly from disk with Python/openpyxl."
  • "The file exists at ...."
  • "The workbook contains sheets ...."
  • "The verification checklist found N pass items, M fail items, and K unavailable-in-source items."
  • "Because one mandatory criterion failed, I am not declaring completion."
  • "Because all mandatory criteria passed, the workbook is ready."

Anti-patterns to avoid

Do not:

  • finalize because an agent said it wrote the file
  • rely only on workbook-edit success messages
  • treat category coverage as satisfied without checking categories directly
  • assume sheet names from memory
  • quote counts without direct disk inspection
  • ignore a failed verification line
  • downgrade a real failure into vague prose
  • call a task complete when proof shows mandatory gaps

Why this skill exists

Spreadsheet tasks often fail not during editing but during final judgment. A workbook can be modified successfully and still be wrong.

This skill prevents false completion by forcing:

  • anchored workbook identity checks
  • deterministic post-write inspection
  • criterion-by-criterion proof
  • a hard stop when mandatory requirements remain unsatisfied

When workbook correctness matters, proof beats summary.

© HKUDS, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file in benchmarks/gdpval/skills/spreadsheet-audit-edit-validate-merged of HKUDS/OpenSpace.

  • SKILL.md
  • .skill_id

Open the folder on GitHubat commit 3827781

Compare with similar skills

Spreadsheet Proof Gate next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Spreadsheet Proof Gate compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Spreadsheet Proof Gate this skillHKUDS/OpenSpace7.7k—~4.5kAutomated safety check: PassMIT
Excel Workbook EditorTokenRhythm/opensquilla7.1k—~1.6kAutomated safety check: PassApache-2.0
Kimi XLSXthvroyal/kimi-skills238—~9.5kAutomated safety check: PassNone
Office XLSXopen-octo/octo-agent125—~1.8kAutomated safety check: NotesMIT
Cc Streaming Export Safetydoccker/cc-use-exp1.1k—~2.2kAutomated safety check: PassCustom licence
XLSXnexus-research-lab/nexus150—~501Automated safety check: PassApache-2.0

Similar skills

  • Excel Workbook Editor

    TokenRhythm/opensquilla

    Inspects, edits in place or creates Microsoft Excel .xlsx workbooks with openpyxl, treating each cell as a typed number, string, datetime or formula value.

    7.1k GitHub stars~1.6k tokensUpdated 3 days ago
    Documents & OfficeAuto-check passed

  • Kimi XLSX

    thvroyal/kimi-skills

    Specialized utility for advanced manipulation, analysis, and creation of spreadsheet files, including (but not limited to) XLSX, XLSM, CSV formats.

    238 GitHub stars~9.5k tokensUpdated 6 mo ago
    Documents & OfficeAuto-check passed

  • Office XLSX

    open-octo/octo-agent

    Create, read, and edit Excel (.xlsx) spreadsheets programmatically with openpyxl — cell values, formulas, styling (fonts/fills/borders/alignment/number formats), merged cells, multiple sheets…

    125 GitHub stars~1.8k tokensUpdated today
    Documents & OfficeAuto-check: notes

  • Cc Streaming Export Safety

    doccker/cc-use-exp

    当实现用户驱动的大文件导出或批量序列化（Excel/CSV/JSON/JSONL/PDF，数据量未知或超过 1 万行/10 MB）时触发；普通小文件下载、静态资源下载、非导出 Writer/Report 类不触发。防止 OOM、临时文件残留、同步导出阻塞 HTTP 线程和表格公式注入。

    1.1k GitHub stars~2.2k tokensUpdated 1 mo ago
    Documents & OfficeAuto-check passed

  • XLSX

    nexus-research-lab/nexus

    只要电子表格文件（.xlsx、.xlsm、旧版 .xls、.csv、.tsv）是任务的主要输入或输出，或者需要 被打开、读取、创建、修改，就使用本 Skill。也适用于公式、格式、图表、数据清洗和表格格式转换。

    150 GitHub stars~501 tokensUpdated today
    Documents & OfficeAuto-check passed

  • Office XLSX

    singula-ai/alego

    Read, create, and modify Excel workbooks (.xlsx), including data, formulas, formatting, and pandas analysis.

    109 GitHub starsUsed in 1 repo~2.3k tokens
    Documents & OfficeAuto-check passed

More from HKUDS/OpenSpace

All 199 skills in this repo

  • Audio Track Production Workflow

    HKUDS/OpenSpace

    Walks through producing a master audio track plus stems in Python, from checking a reference file and timing sections by BPM to effects, a zip archive and final verification.

    7.7k GitHub stars~2.9k tokensUpdated 1 mo ago
    Auto-check passed

  • Cascade Failure Recovery

    HKUDS/OpenSpace

    Handle cascading data retrieval tool failures by falling back to embedded knowledge generation

    7.7k GitHub stars~765 tokensUpdated 1 mo ago
    Auto-check passed

  • Sandbox Code Execution Fallback

    HKUDS/OpenSpace

    Gives an agent a workaround when its code-execution sandbox keeps failing: save the Python script to a file and run it through the shell instead.

    7.7k GitHub stars~588 tokensUpdated 1 mo ago
    Auto-check passed

  • Code Execution Fallback

    HKUDS/OpenSpace

    A recovery routine for agents whose sandboxed code runner keeps failing: save the Python script to disk, then run it through the shell and read the output.

    7.7k GitHub stars~652 tokensUpdated 1 mo ago
    Auto-check passed

  • Code Execution Fallback and Path Anchoring

    HKUDS/OpenSpace

    Fallback ladder for failed sandboxed code runs, plus the habit of fixing the working directory first so generated files land in the right place.

    7.7k GitHub stars~1.1k tokensUpdated 1 mo ago
    Auto-check passed

  • Code Execution Fallback E81068

    HKUDS/OpenSpace

    Fallback workflow for executing Python code when executecodesandbox fails repeatedly

    7.7k GitHub stars~1.1k tokensUpdated 1 mo ago
    Auto-check passed

Works with

Categories

Questions about Spreadsheet Proof Gate

What does Spreadsheet Proof Gate do?

Edit Excel workbooks only after anchored preflight audit, then require deterministic Python/openpyxl post-write proof that every requested criterion is satisfied or explicitly reconciled before…. Spreadsheet Proof Gate is an agent skill from HKUDS/OpenSpace. Edit Excel workbooks only after anchored preflight audit, then require deterministic Python/openpyxl post-write proof that every requested criterion is satisfied or explicitly reconciled before finalizing.

When should I use Spreadsheet Proof Gate?

Spreadsheet Proof Gate fits situations like: tasks that involve Excel spreadsheets.

How do I install Spreadsheet Proof Gate in Claude Code?

Run `npx skills add HKUDS/OpenSpace --skill spreadsheet-proof-gate -a claude-code`. Or copy the skill folder (benchmarks/gdpval/skills/spreadsheet-audit-edit-validate-merged in HKUDS/OpenSpace) into .claude/skills/spreadsheet-proof-gate in your project. Claude Code loads it when a task matches its description.

How do I install Spreadsheet Proof Gate in Codex?

Run `npx skills add HKUDS/OpenSpace --skill spreadsheet-proof-gate -a codex`. Or copy the skill folder (benchmarks/gdpval/skills/spreadsheet-audit-edit-validate-merged in HKUDS/OpenSpace) into .agents/skills/spreadsheet-proof-gate in your project. Codex loads it when a task matches its description.

Can I use Spreadsheet Proof Gate in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add HKUDS/OpenSpace --skill spreadsheet-proof-gate -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/spreadsheet-proof-gate, .gemini/skills/spreadsheet-proof-gate, .github/skills/spreadsheet-proof-gate and .opencode/skills/spreadsheet-proof-gate in your project.

What does Spreadsheet Proof Gate need to run?

SKILL.md names no scripts, command-line tools or credentials: Spreadsheet Proof Gate is instructions for the agent only. Our summary lists: Python 3.

Does Spreadsheet Proof Gate access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Spreadsheet Proof Gate safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Spreadsheet Proof Gate use?

Spreadsheet Proof Gate is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Spreadsheet Proof Gate use?

About 4.5k tokens (SKILL.md is roughly 18k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Spreadsheet Proof Gate?

Skills that share tags, products or a category with Spreadsheet Proof Gate: Excel Workbook Editor (TokenRhythm/opensquilla, 7.1k stars), Kimi XLSX (thvroyal/kimi-skills, 238 stars), Office XLSX (open-octo/octo-agent, 125 stars) and Cc Streaming Export Safety (doccker/cc-use-exp, 1.1k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Spreadsheet Proof Gate?

HKUDS (a GitHub organization) maintains it in HKUDS/OpenSpace, which has 7,749 GitHub stars. The repository holds 199 skills in this directory. The repository was last updated on August 12, 2026.

Source: HKUDS/OpenSpace on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.