[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6KG1RENXBV8GPQ1R2A5BT76":3,"skill-same:01M6KG1RENXBV8GPQ1R2A5BT76":68},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":35,"linkOnly":36,"markdownContent":37,"markdownBody":38,"frontMatter":39,"refName":40,"commitSha":41,"skillDirectory":42,"forksCount":14,"repositoryHtmlUrl":43,"repositoryHomepageUrl":6,"ownerHtmlUrl":44,"ownerBlogUrl":45,"ownerDescription":46,"updatedAt":47,"install":48,"resources":61},"01M6KG1RENXBV8GPQ1R2A5BT76","sandbox-file-discovery-and-validation",null,"Use shell tools safely in sandboxed environments by discovering files from real workspace roots instead of assumed user folders, then explicitly verifying output artifacts after generation.","xyliu-cs\u002FOpenSpace-GDPval","xyliu-cs","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F89236536?v=4","01M6KF8Z1V4TXS86CKSH48BY8T","xy","01M6KF8Z67443W0SHXM85JEF0B",0,"MIT","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fsandbox-file-discovery-and-validation\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34],"sandbox","validation","file","discovery","file discovery","sandbox file discovery and validation","artifacts","Tools","shell","workspace","user","use","files","generation","output","environments","2026-11-10 10:12:30",false,"---\nname: sandbox-file-discovery-and-validation\ndescription: Use shell tools safely in sandboxed environments by discovering files from real workspace roots instead of assumed user folders, then explicitly verifying output artifacts after generation.\n---\n\n# Sandbox File Discovery and Validation\n\nUse this skill when working in constrained or sandboxed shell environments where common folders like `~\u002FDesktop`, `~\u002FDocuments`, or `~\u002FDownloads` may not exist, may be inaccessible, or may not be where task files are stored.\n\nThe goal is to:\n1. discover files from actual accessible roots,\n2. avoid brittle path assumptions,\n3. generate outputs in known locations, and\n4. explicitly validate that outputs were created and match expectations.\n\n## Core principles\n\n- Do not assume GUI-era user folders exist.\n- Start from the current working directory and other confirmed workspace roots.\n- Use recursive search from known roots, not from `\u002F` unless necessary.\n- Prefer deterministic output paths you control.\n- After creating an artifact, verify it exists, is readable, and matches the requested form.\n\n## When to use this\n\nUse this pattern when:\n- you need to locate input files in an unfamiliar sandbox,\n- the environment may be ephemeral or nonstandard,\n- you are producing files for the user,\n- success depends on the actual contents or structure of the output.\n\n## Recommended workflow\n\n### 1) Establish your real working roots\n\nBegin by identifying where you are and what directories are actually available.\n\nExample:\n`pwd`\n`ls -la`\n`find . -maxdepth 2 -type d | sort`\n\nIf needed, inspect nearby likely roots:\n`ls -la \u002Ftmp`\n`ls -la \u002Fworkspace 2>\u002Fdev\u002Fnull || true`\n`ls -la \u002Fworkspaces 2>\u002Fdev\u002Fnull || true`\n`ls -la \u002Fmnt\u002Fdata 2>\u002Fdev\u002Fnull || true`\n\nTreat only confirmed, readable directories as search roots.\n\n## 2) Avoid assumed folders\n\nDo not begin with paths like:\n- `~\u002FDesktop`\n- `~\u002FDocuments`\n- `~\u002FDownloads`\n\nunless you have already confirmed they exist and are relevant.\n\nBad:\n`find ~\u002FDesktop -name \"*.xlsx\"`\n\nBetter:\n`find . -type f -name \"*.xlsx\"`\nor, if a root is confirmed:\n`find \u002Fworkspace -type f -name \"*.xlsx\" 2>\u002Fdev\u002Fnull`\n\n## 3) Search from known roots with bounded, focused queries\n\nPrefer targeted searches over broad filesystem scans.\n\nUseful patterns:\n`find . -type f | sort`\n`find . -type f -iname \"*report*\"`\n`find . -type f \\( -iname \"*.csv\" -o -iname \"*.xlsx\" -o -iname \"*.json\" \\)`\n\nIf multiple roots are known:\n`find . \u002Fworkspace \u002Fmnt\u002Fdata -type f 2>\u002Fdev\u002Fnull | sort`\n\nGuidelines:\n- suppress permission noise with `2>\u002Fdev\u002Fnull` when appropriate,\n- filter by extension or name fragment,\n- sort output for stable inspection,\n- keep searches bounded to known roots.\n\n## 4) Choose a deterministic output location\n\nWhen generating a file, write it somewhere explicit and easy to re-check.\n\nGood patterns:\n- current directory,\n- a task-specific subdirectory like `.\u002Foutput`,\n- a confirmed writable workspace directory.\n\nExample:\n`mkdir -p output`\n`python make_result.py --out output\u002Ffinal.xlsx`\n\nAvoid writing to guessed locations that may not exist.\n\n## 5) Immediately validate the output artifact\n\nNever assume generation succeeded just because a command exited successfully.\n\nAt minimum, confirm:\n- the file exists,\n- the path is the one you intended,\n- the file is non-empty when applicable,\n- the format or structure matches the task.\n\nBasic checks:\n`ls -l output\u002Ffinal.xlsx`\n`test -f output\u002Ffinal.xlsx && echo \"exists\"`\n`test -s output\u002Ffinal.xlsx && echo \"non-empty\"`\n`file output\u002Ffinal.xlsx`\n\nFor text-like outputs:\n`head -n 20 output\u002Fresult.csv`\n`wc -l output\u002Fresult.csv`\n\nFor structured outputs, inspect with the relevant tool:\n- CSV: preview headers and row counts\n- JSON: parse and inspect keys\n- ZIP-like formats such as `.xlsx`: verify internal structure or open with a library\n- generated code\u002Fconfig: run syntax or schema checks if available\n\n## 6) Validate against task requirements, not just file existence\n\nA file existing is necessary but not sufficient.\n\nCheck the artifact against the requested deliverable:\n- expected filename or location,\n- expected sheet names, tabs, columns, or keys,\n- expected formulas, calculations, or derived values,\n- expected number of records,\n- expected transformations or formatting.\n\nExamples:\n- If a workbook was requested, verify the workbook contains the required sheets.\n- If formulas were required, verify formulas are present, not just pasted values.\n- If a filtered dataset was requested, verify row counts and selected columns.\n- If a renamed file was requested, verify the actual output name.\n\n## 7) Report with evidence\n\nWhen you finish, cite the exact artifact path and the validation you performed.\n\nGood completion style:\n- “Created `output\u002Ffinal.xlsx`.”\n- “Verified it exists and is non-empty.”\n- “Confirmed workbook contains sheets `Summary` and `Data`.”\n- “Checked that computed totals are present in column F.”\n\nThis reduces false positives where an output was produced but not actually checked.\n\n# Shell patterns to reuse\n\n## Minimal discovery sequence\n\n`pwd`\n`ls -la`\n`find . -maxdepth 3 -type f | sort`\n\n## Targeted file search\n\n`find . -type f \\( -iname \"*.xlsx\" -o -iname \"*.csv\" -o -iname \"*.tsv\" \\) | sort`\n\n## Multi-root search with graceful fallback\n\n`find . \u002Fworkspace \u002Fmnt\u002Fdata -type f 2>\u002Fdev\u002Fnull | sort`\n\n## Output creation and validation\n\n`mkdir -p output`\n`some_command > output\u002Fresult.txt`\n`test -f output\u002Fresult.txt && test -s output\u002Fresult.txt && echo \"validated\"`\n\n# Practical decision rules\n\n## Prefer this\n- `pwd` to anchor yourself\n- `find` from `.` or other verified roots\n- explicit output directories\n- post-generation checks tied to the requested deliverable\n\n## Avoid this\n- guessing user-centric folders\n- searching the entire filesystem first\n- declaring success after creation without inspection\n- reporting only the command you ran instead of what you verified\n\n# Example end-to-end pattern\n\n1. Discover roots:\n   `pwd && ls -la`\n\n2. Find candidate inputs:\n   `find . -type f \\( -iname \"*.xlsx\" -o -iname \"*.csv\" \\) | sort`\n\n3. Create controlled output location:\n   `mkdir -p output`\n\n4. Generate artifact:\n   `python transform.py input\u002Fsource.csv output\u002Ffinal.csv`\n\n5. Validate artifact exists and inspect content:\n   `ls -l output\u002Ffinal.csv`\n   `test -s output\u002Ffinal.csv`\n   `head -n 5 output\u002Ffinal.csv`\n\n6. Validate requirement-specific properties:\n   - expected headers present,\n   - row count reasonable,\n   - transformations applied.\n\n# Success criteria\n\nThis pattern is applied correctly when:\n- file discovery starts from confirmed accessible roots,\n- no critical step depends on unverified assumed directories,\n- generated artifacts are written to known locations,\n- outputs are explicitly inspected or checked,\n- final reporting includes evidence of validation.","# Sandbox File Discovery and Validation\n\nUse this skill when working in constrained or sandboxed shell environments where common folders like `~\u002FDesktop`, `~\u002FDocuments`, or `~\u002FDownloads` may not exist, may be inaccessible, or may not be where task files are stored.\n\nThe goal is to:\n1. discover files from actual accessible roots,\n2. avoid brittle path assumptions,\n3. generate outputs in known locations, and\n4. explicitly validate that outputs were created and match expectations.\n\n## Core principles\n\n- Do not assume GUI-era user folders exist.\n- Start from the current working directory and other confirmed workspace roots.\n- Use recursive search from known roots, not from `\u002F` unless necessary.\n- Prefer deterministic output paths you control.\n- After creating an artifact, verify it exists, is readable, and matches the requested form.\n\n## When to use this\n\nUse this pattern when:\n- you need to locate input files in an unfamiliar sandbox,\n- the environment may be ephemeral or nonstandard,\n- you are producing files for the user,\n- success depends on the actual contents or structure of the output.\n\n## Recommended workflow\n\n### 1) Establish your real working roots\n\nBegin by identifying where you are and what directories are actually available.\n\nExample:\n`pwd`\n`ls -la`\n`find . -maxdepth 2 -type d | sort`\n\nIf needed, inspect nearby likely roots:\n`ls -la \u002Ftmp`\n`ls -la \u002Fworkspace 2>\u002Fdev\u002Fnull || true`\n`ls -la \u002Fworkspaces 2>\u002Fdev\u002Fnull || true`\n`ls -la \u002Fmnt\u002Fdata 2>\u002Fdev\u002Fnull || true`\n\nTreat only confirmed, readable directories as search roots.\n\n## 2) Avoid assumed folders\n\nDo not begin with paths like:\n- `~\u002FDesktop`\n- `~\u002FDocuments`\n- `~\u002FDownloads`\n\nunless you have already confirmed they exist and are relevant.\n\nBad:\n`find ~\u002FDesktop -name \"*.xlsx\"`\n\nBetter:\n`find . -type f -name \"*.xlsx\"`\nor, if a root is confirmed:\n`find \u002Fworkspace -type f -name \"*.xlsx\" 2>\u002Fdev\u002Fnull`\n\n## 3) Search from known roots with bounded, focused queries\n\nPrefer targeted searches over broad filesystem scans.\n\nUseful patterns:\n`find . -type f | sort`\n`find . -type f -iname \"*report*\"`\n`find . -type f \\( -iname \"*.csv\" -o -iname \"*.xlsx\" -o -iname \"*.json\" \\)`\n\nIf multiple roots are known:\n`find . \u002Fworkspace \u002Fmnt\u002Fdata -type f 2>\u002Fdev\u002Fnull | sort`\n\nGuidelines:\n- suppress permission noise with `2>\u002Fdev\u002Fnull` when appropriate,\n- filter by extension or name fragment,\n- sort output for stable inspection,\n- keep searches bounded to known roots.\n\n## 4) Choose a deterministic output location\n\nWhen generating a file, write it somewhere explicit and easy to re-check.\n\nGood patterns:\n- current directory,\n- a task-specific subdirectory like `.\u002Foutput`,\n- a confirmed writable workspace directory.\n\nExample:\n`mkdir -p output`\n`python make_result.py --out output\u002Ffinal.xlsx`\n\nAvoid writing to guessed locations that may not exist.\n\n## 5) Immediately validate the output artifact\n\nNever assume generation succeeded just because a command exited successfully.\n\nAt minimum, confirm:\n- the file exists,\n- the path is the one you intended,\n- the file is non-empty when applicable,\n- the format or structure matches the task.\n\nBasic checks:\n`ls -l output\u002Ffinal.xlsx`\n`test -f output\u002Ffinal.xlsx && echo \"exists\"`\n`test -s output\u002Ffinal.xlsx && echo \"non-empty\"`\n`file output\u002Ffinal.xlsx`\n\nFor text-like outputs:\n`head -n 20 output\u002Fresult.csv`\n`wc -l output\u002Fresult.csv`\n\nFor structured outputs, inspect with the relevant tool:\n- CSV: preview headers and row counts\n- JSON: parse and inspect keys\n- ZIP-like formats such as `.xlsx`: verify internal structure or open with a library\n- generated code\u002Fconfig: run syntax or schema checks if available\n\n## 6) Validate against task requirements, not just file existence\n\nA file existing is necessary but not sufficient.\n\nCheck the artifact against the requested deliverable:\n- expected filename or location,\n- expected sheet names, tabs, columns, or keys,\n- expected formulas, calculations, or derived values,\n- expected number of records,\n- expected transformations or formatting.\n\nExamples:\n- If a workbook was requested, verify the workbook contains the required sheets.\n- If formulas were required, verify formulas are present, not just pasted values.\n- If a filtered dataset was requested, verify row counts and selected columns.\n- If a renamed file was requested, verify the actual output name.\n\n## 7) Report with evidence\n\nWhen you finish, cite the exact artifact path and the validation you performed.\n\nGood completion style:\n- “Created `output\u002Ffinal.xlsx`.”\n- “Verified it exists and is non-empty.”\n- “Confirmed workbook contains sheets `Summary` and `Data`.”\n- “Checked that computed totals are present in column F.”\n\nThis reduces false positives where an output was produced but not actually checked.\n\n# Shell patterns to reuse\n\n## Minimal discovery sequence\n\n`pwd`\n`ls -la`\n`find . -maxdepth 3 -type f | sort`\n\n## Targeted file search\n\n`find . -type f \\( -iname \"*.xlsx\" -o -iname \"*.csv\" -o -iname \"*.tsv\" \\) | sort`\n\n## Multi-root search with graceful fallback\n\n`find . \u002Fworkspace \u002Fmnt\u002Fdata -type f 2>\u002Fdev\u002Fnull | sort`\n\n## Output creation and validation\n\n`mkdir -p output`\n`some_command > output\u002Fresult.txt`\n`test -f output\u002Fresult.txt && test -s output\u002Fresult.txt && echo \"validated\"`\n\n# Practical decision rules\n\n## Prefer this\n- `pwd` to anchor yourself\n- `find` from `.` or other verified roots\n- explicit output directories\n- post-generation checks tied to the requested deliverable\n\n## Avoid this\n- guessing user-centric folders\n- searching the entire filesystem first\n- declaring success after creation without inspection\n- reporting only the command you ran instead of what you verified\n\n# Example end-to-end pattern\n\n1. Discover roots:\n   `pwd && ls -la`\n\n2. Find candidate inputs:\n   `find . -type f \\( -iname \"*.xlsx\" -o -iname \"*.csv\" \\) | sort`\n\n3. Create controlled output location:\n   `mkdir -p output`\n\n4. Generate artifact:\n   `python transform.py input\u002Fsource.csv output\u002Ffinal.csv`\n\n5. Validate artifact exists and inspect content:\n   `ls -l output\u002Ffinal.csv`\n   `test -s output\u002Ffinal.csv`\n   `head -n 5 output\u002Ffinal.csv`\n\n6. Validate requirement-specific properties:\n   - expected headers present,\n   - row count reasonable,\n   - transformations applied.\n\n# Success criteria\n\nThis pattern is applied correctly when:\n- file discovery starts from confirmed accessible roots,\n- no critical step depends on unverified assumed directories,\n- generated artifacts are written to known locations,\n- outputs are explicitly inspected or checked,\n- final reporting includes evidence of validation.",{"name":5,"description":7},"main","af1b40a42167d09c30b3a63000aa14ace4c1e229","gdpval_bench\u002Fskills\u002Fsandbox-file-discovery-and-validation","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval","https:\u002F\u002Fgithub.com\u002Fxyliu-cs","https:\u002F\u002Fxyliu-cs.github.io","PhD Student in CS, CUHK-Shenzhen.","2026-04-06 20:32:15",{"source":8,"skillName":5,"commands":49,"prompt":59,"zipAvailable":60},[50,53,56],{"tool":51,"command":52},"npx","npx skills add xyliu-cs\u002FOpenSpace-GDPval --skill sandbox-file-discovery-and-validation",{"tool":54,"command":55},"bunx","bunx skills add xyliu-cs\u002FOpenSpace-GDPval --skill sandbox-file-discovery-and-validation",{"tool":57,"command":58},"pnpm","pnpm dlx skills add xyliu-cs\u002FOpenSpace-GDPval --skill sandbox-file-discovery-and-validation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fsandbox-file-discovery-and-validation\u002FSKILL.md",true,[62],{"path":63,"resourceType":64,"resourceTypeDesc":65,"sizeBytes":66,"sourceHtmlUrl":67},".skill_id","OTHER","其他",51,"https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fsandbox-file-discovery-and-validation\u002F.skill_id",[69,82,90,106,131,152],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":35},"01M6KG8MF1S664SWBDENYRSCK8","zip-export-deliverable","Ensures codebase deliverables are properly exported as ZIP archives through a dedicated finalization step","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fzip-export-deliverable\u002FSKILL.md",[],[76,77,78,79,80,81],"zip export deliverable","zip","export","deliverables","codebase","as",{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":35},"01M6KG8GAZWHHEK7VNJDW0ME21","zip-deliverable-finalization","Ensure codebase deliverables are properly packaged as ZIP exports in a dedicated final step","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fzip-deliverable-finalization\u002FSKILL.md",[],[89,77,79,80,81],"zip deliverable finalization",{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":35},"01M6KG8CMQQ1PYMNTSRZ9VS1MD","zero-iteration-failure-analysis","Identify and handle agent failures with 0 iterations as pre-execution system issues","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fzero-iteration-failure-analysis\u002FSKILL.md",[],[97,98,99,100,101,102,103,104,81,105],"failure analysis","ZeRO","analysis","iteration","zero iteration failure analysis","with","issues","execution","system",{"skillNo":107,"name":108,"nameZh":6,"description":109,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":117,"publishedAt":35},"01M6KG86W3PA07FT9W58YHHERM","write-file-fallback-report","Generate professional documents using write_file when primary data sources fail, leveraging embedded domain knowledge instead of external retrieval","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fwrite-file-fallback-report\u002FSKILL.md",[112],{"categoryNo":113,"categoryName":114,"taxonomyType":115,"taxonomyTypeDesc":116},"01JBSEEDCAT000000000000034","行政与办公人员","PROFESSION","适用职业",[118,119,21,120,121,122,123,124,125,126,127,128,129,130],"fallback","write","report","write file fallback report","domain","professional","sources","retrieval","generate","embedded","documents","data","knowledge",{"skillNo":132,"name":133,"nameZh":6,"description":134,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":135,"categories":136,"keywords":142,"publishedAt":35},"01M6KG82GAEY3GG77CMGCK1C2T","fallback-doc-with-delegation","Generate professional documents after tool failures using shell_agent for complex formats or write_file for simple text\u002Fmarkdown","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fwrite-file-fallback-report-enhanced\u002FSKILL.md",[137],{"categoryNo":138,"categoryName":139,"taxonomyType":140,"taxonomyTypeDesc":141},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[143,102,118,144,145,27,128,126,21,146,147,148,119,149,123,150,151],"fallback doc with delegation","doc","delegation","markdown","tool","text","simple","formats","for",{"skillNo":153,"name":154,"nameZh":6,"description":155,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":156,"categories":157,"keywords":159,"publishedAt":35},"01M6KG7WT7GX5NYCGEFHW22NV7","context-anchored-fallback-report","Generate documents with write_file when retrieval tools fail, with explicit guardrails against task context drift","https:\u002F\u002Fgithub.com\u002Fxyliu-cs\u002FOpenSpace-GDPval\u002Fblob\u002Fmain\u002Fgdpval_bench\u002Fskills\u002Fwrite-file-fallback-report-enhanced-4bc5c2\u002FSKILL.md",[158],{"categoryNo":113,"categoryName":114,"taxonomyType":115,"taxonomyTypeDesc":116},[118,120,160,161,126,21,162,26,119,163,164,125,102,165,128],"context","context anchored fallback report","Guardrails","task","drift","context drift"]