[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79N9QF63TTY3B3XB5VHEC48":3,"skill-same:01M79N9QF63TTY3B3XB5VHEC48":47},{"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":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":14,"repositoryHtmlUrl":29,"repositoryHomepageUrl":6,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M79N9QF63TTY3B3XB5VHEC48","config-gc",null,"Garbage collection for your Claude Code configuration. Periodically scans ~\u002F.claude (skills, memory, hooks, permissions, MCP servers, caches) for redundant, stale, orphaned, or low-value items, then walks the user through a confirm-each-deletion cleanup. Use when the user says \"clean up my config\", \"config GC\", \"too many skills\", \"audit my setup\", \"my .claude is bloated\", or asks for a periodic config review.","JunMystery\u002FAgent-Guidance","JunMystery","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F161468934?v=4","01M79MTMV8F0BY8NK1A92RA8A0","Jun","01M79MTMWCFBKPHN5B25E6V20B",0,"MIT","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fconfig-gc\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: config-gc\ndescription: Garbage collection for your Claude Code configuration. Periodically scans ~\u002F.claude (skills, memory, hooks, permissions, MCP servers, caches) for redundant, stale, orphaned, or low-value items, then walks the user through a confirm-each-deletion cleanup. Use when the user says \"clean up my config\", \"config GC\", \"too many skills\", \"audit my setup\", \"my .claude is bloated\", or asks for a periodic config review.\nmetadata:\n  origin: ECC\n---\n\n# Config GC — Garbage Collection for Claude Code Setups\n\nBorrowed from runtime garbage collection: periodically scan for objects that are no longer referenced, redundant, expired, or low-value, and reclaim the space. The critical difference: **here, collection requires a human in the loop. Never delete autonomously.**\n\n## When to Activate\n\n- The user asks to clean up, audit, or slim down their Claude Code configuration\n- The user complains about too many skills, noisy hooks, or slow session startup\n- A monthly\u002Fperiodic config review is due\n- After installing a large skill pack (e.g. this repo), to reconcile overlaps with existing setup\n\nDo NOT activate for: cleaning project source code (that's refactoring), clearing chat history, or uninstalling Claude Code itself.\n\n## Design Philosophy\n\n1. **Append-only configs leak.** Skills, memory files, hooks, and permission entries only ever get added. Without periodic review they rot silently.\n2. **Regular audits beat one-time purges.** Scan every ~30 days, propose a small batch of candidates each time.\n3. **Per-channel strategies.** Each accumulation type (skills, hooks, permissions, ...) has its own staleness signals — don't apply one rule everywhere.\n4. **Soft-delete first.** Rename to `.disabled` > move to `~\u002F.claude\u002F_gc_trash\u002F` > real deletion. Always keep an undo path.\n5. **Forced human-in-the-loop.** Every candidate gets its own `[y\u002Fn\u002Fskip]` confirmation. No \"yes to all\" shortcut.\n6. **Keep a log.** Every GC run appends to `~\u002F.claude\u002Fgc_log.md`: what was touched, why, and how to undo it.\n\n## Scan Channels\n\n| # | Channel | Path | Staleness \u002F redundancy signals |\n|---|---------|------|--------------------------------|\n| 1 | Skills | `~\u002F.claude\u002Fskills\u002F*\u002F` | Heavily overlapping names; never triggered in recent transcripts; domain mismatch with the user's actual work; broken or empty SKILL.md |\n| 2 | Memory | `~\u002F.claude\u002F**\u002Fmemory\u002F*.md` + its index | Multiple index entries for one topic; contents contradicting newer entries; dates that have passed; orphan files missing from the index; sub-100-word fragments that should merge |\n| 3 | Hooks | `~\u002F.claude\u002Fhooks\u002F` + settings | Scripts present on disk but referenced by no hook config; old versions superseded by rewrites |\n| 4 | Permissions | `permissions.allow` in `settings.json` \u002F `settings.local.json` | Duplicate entries; specific entries already covered by a wildcard (e.g. `Bash(git push)` when `Bash(*)` is allowed); one-off grants from past experiments |\n| 5 | MCP servers | `~\u002F.claude.json` or project `.mcp.json` | Servers that fail to connect; functional duplicates; long-unused |\n| 6 | Scheduled reminders \u002F jobs | wherever the user keeps them | Fired one-shots older than 30 days; jobs whose target scripts no longer exist |\n| 7 | Project history | `~\u002F.claude\u002Fprojects\u002F*\u002F` | Stale handoff snapshots; session records superseded by newer state |\n| 8 | Runtime caches | `cache\u002F`, `file-history\u002F`, `logs\u002F`, `shell-snapshots\u002F` | Sort by size and mtime; propose items >30 days old and large |\n\n## Workflow\n\n1. **Scan** all channels (or the subset the user names). Collect candidates with: path, channel, signal that flagged it, size, last-modified.\n2. **Rank** by confidence (broken\u002Forphaned = high; merely old = low) and present as a numbered table. Cap each run at ~20 candidates — GC is periodic, not exhaustive.\n3. **Confirm one by one.** For each candidate show the evidence, then ask `[y\u002Fn\u002Fskip]`. The user can stop at any point.\n4. **Soft-delete confirmed items**: prefer `.disabled` rename for skills\u002Fhooks and `_gc_trash\u002F\u003Cdate>\u002F` move for files. Permission entries live in JSON (no comments possible): back up the settings file, record each removed entry verbatim in `gc_log.md`, then remove it from the `allow` array with `jq`. Only hard-delete when the user explicitly asks.\n5. **Log** the run to `~\u002F.claude\u002Fgc_log.md`: timestamp, items actioned, undo instructions.\n6. **Report**: reclaimed size, channels still healthy, suggested next review date.\n\n## Example Scan Commands\n\nOrphaned hook scripts (channel 3) — scripts on disk that no hook config references:\n\n```bash\nfor f in ~\u002F.claude\u002Fhooks\u002F*; do\n  name=$(basename \"$f\")\n  grep -rq \"$name\" ~\u002F.claude\u002Fsettings.json ~\u002F.claude\u002Fsettings.local.json 2>\u002Fdev\u002Fnull \\\n    || echo \"ORPHAN: $f\"\ndone\n```\n\nRedundant permission entries (channel 4) — duplicates, and specific grants shadowed by a wildcard:\n\n```bash\njq -r '.permissions.allow[]' ~\u002F.claude\u002Fsettings.local.json | sort | uniq -d\nif jq -e '.permissions.allow | index(\"Bash(*)\")' ~\u002F.claude\u002Fsettings.local.json >\u002Fdev\u002Fnull; then\n  jq -r '.permissions.allow[]' ~\u002F.claude\u002Fsettings.local.json \\\n    | grep '^Bash(' | grep -vF 'Bash(*)'\nfi\n```\n\nLargest stale caches (channel 8) — `du -k` instead of GNU-only `find -printf`, so it works on macOS\u002FBSD too:\n\n```bash\nfind ~\u002F.claude\u002Ffile-history ~\u002F.claude\u002Fshell-snapshots -type f -mtime +30 \\\n  -exec du -k {} + 2>\u002Fdev\u002Fnull | sort -rn | head -20\n```\n\nSoft-delete with undo path (capture the date once so the log can't disagree with the directory):\n\n```bash\ngc_date=$(date +%Y-%m-%d)\nmkdir -p ~\u002F.claude\u002F_gc_trash\u002F$gc_date\nmv ~\u002F.claude\u002Fskills\u002Fdead-skill ~\u002F.claude\u002F_gc_trash\u002F$gc_date\u002F\necho \"$(date -Iseconds) moved skills\u002Fdead-skill -> _gc_trash\u002F$gc_date\u002F (undo: mv back)\" >> ~\u002F.claude\u002Fgc_log.md\n```\n\nRemoving a confirmed-redundant permission entry (JSON has no comments — back up, log, then edit):\n\n```bash\ncp ~\u002F.claude\u002Fsettings.local.json ~\u002F.claude\u002Fsettings.local.json.bak\necho \"$(date -Iseconds) removed permission entry: Bash(git push) (undo: restore from .bak or re-add)\" >> ~\u002F.claude\u002Fgc_log.md\njq '.permissions.allow -= [\"Bash(git push)\"]' ~\u002F.claude\u002Fsettings.local.json.bak \\\n  > ~\u002F.claude\u002Fsettings.local.json\n```\n\n## Anti-Patterns\n\n- **Bulk approval.** Asking \"delete all 15? [y\u002Fn]\" defeats the design. One item, one decision.\n- **Hard-deleting on first pass.** If there's no `_gc_trash\u002F` copy or `.disabled` rename, you did it wrong.\n- **Treating \"old\" as \"dead\".** A skill untouched for 60 days may be seasonal (tax season, quarterly reviews). Age is a signal, not a verdict — that's why a human confirms.\n- **Cleaning memory by truncation.** Merging two contradicting memory files requires reading both and keeping the newer truth, not deleting the longer one.\n- **Touching anything outside `~\u002F.claude`** (or the project's `.claude\u002F`). Config GC never wanders into source trees.\n\n## Best Practices\n\n- Run after big additions, not just on a calendar: installing a 50-skill pack is exactly when overlap with existing skills appears.\n- When two skills overlap, prefer disabling the one with the weaker trigger description — it's the one that was probably never firing anyway.\n- Permission cleanup is the highest-value channel per minute spent: redundant allow-entries make security review harder.\n- Keep `gc_log.md` forever. It's tiny, and \"when did I disable that hook and why\" comes up more often than you'd think.\n\n## Related Skills\n\n- `skill-stocktake` — audits skill *quality*; config-gc audits skill *existence*. Run stocktake on what survives GC.\n- `workspace-surface-audit` — the additive counterpart: recommends what to install. config-gc is the subtractive half of the same lifecycle.\n- `configure-ecc` — after installing skills with it, run config-gc to reconcile overlaps with your pre-existing setup.\n- `continuous-learning` — produces the memory files this skill later audits.\n- `security-review` — pairs well with the permissions channel.\n","# Config GC — Garbage Collection for Claude Code Setups\n\nBorrowed from runtime garbage collection: periodically scan for objects that are no longer referenced, redundant, expired, or low-value, and reclaim the space. The critical difference: **here, collection requires a human in the loop. Never delete autonomously.**\n\n## When to Activate\n\n- The user asks to clean up, audit, or slim down their Claude Code configuration\n- The user complains about too many skills, noisy hooks, or slow session startup\n- A monthly\u002Fperiodic config review is due\n- After installing a large skill pack (e.g. this repo), to reconcile overlaps with existing setup\n\nDo NOT activate for: cleaning project source code (that's refactoring), clearing chat history, or uninstalling Claude Code itself.\n\n## Design Philosophy\n\n1. **Append-only configs leak.** Skills, memory files, hooks, and permission entries only ever get added. Without periodic review they rot silently.\n2. **Regular audits beat one-time purges.** Scan every ~30 days, propose a small batch of candidates each time.\n3. **Per-channel strategies.** Each accumulation type (skills, hooks, permissions, ...) has its own staleness signals — don't apply one rule everywhere.\n4. **Soft-delete first.** Rename to `.disabled` > move to `~\u002F.claude\u002F_gc_trash\u002F` > real deletion. Always keep an undo path.\n5. **Forced human-in-the-loop.** Every candidate gets its own `[y\u002Fn\u002Fskip]` confirmation. No \"yes to all\" shortcut.\n6. **Keep a log.** Every GC run appends to `~\u002F.claude\u002Fgc_log.md`: what was touched, why, and how to undo it.\n\n## Scan Channels\n\n| # | Channel | Path | Staleness \u002F redundancy signals |\n|---|---------|------|--------------------------------|\n| 1 | Skills | `~\u002F.claude\u002Fskills\u002F*\u002F` | Heavily overlapping names; never triggered in recent transcripts; domain mismatch with the user's actual work; broken or empty SKILL.md |\n| 2 | Memory | `~\u002F.claude\u002F**\u002Fmemory\u002F*.md` + its index | Multiple index entries for one topic; contents contradicting newer entries; dates that have passed; orphan files missing from the index; sub-100-word fragments that should merge |\n| 3 | Hooks | `~\u002F.claude\u002Fhooks\u002F` + settings | Scripts present on disk but referenced by no hook config; old versions superseded by rewrites |\n| 4 | Permissions | `permissions.allow` in `settings.json` \u002F `settings.local.json` | Duplicate entries; specific entries already covered by a wildcard (e.g. `Bash(git push)` when `Bash(*)` is allowed); one-off grants from past experiments |\n| 5 | MCP servers | `~\u002F.claude.json` or project `.mcp.json` | Servers that fail to connect; functional duplicates; long-unused |\n| 6 | Scheduled reminders \u002F jobs | wherever the user keeps them | Fired one-shots older than 30 days; jobs whose target scripts no longer exist |\n| 7 | Project history | `~\u002F.claude\u002Fprojects\u002F*\u002F` | Stale handoff snapshots; session records superseded by newer state |\n| 8 | Runtime caches | `cache\u002F`, `file-history\u002F`, `logs\u002F`, `shell-snapshots\u002F` | Sort by size and mtime; propose items >30 days old and large |\n\n## Workflow\n\n1. **Scan** all channels (or the subset the user names). Collect candidates with: path, channel, signal that flagged it, size, last-modified.\n2. **Rank** by confidence (broken\u002Forphaned = high; merely old = low) and present as a numbered table. Cap each run at ~20 candidates — GC is periodic, not exhaustive.\n3. **Confirm one by one.** For each candidate show the evidence, then ask `[y\u002Fn\u002Fskip]`. The user can stop at any point.\n4. **Soft-delete confirmed items**: prefer `.disabled` rename for skills\u002Fhooks and `_gc_trash\u002F\u003Cdate>\u002F` move for files. Permission entries live in JSON (no comments possible): back up the settings file, record each removed entry verbatim in `gc_log.md`, then remove it from the `allow` array with `jq`. Only hard-delete when the user explicitly asks.\n5. **Log** the run to `~\u002F.claude\u002Fgc_log.md`: timestamp, items actioned, undo instructions.\n6. **Report**: reclaimed size, channels still healthy, suggested next review date.\n\n## Example Scan Commands\n\nOrphaned hook scripts (channel 3) — scripts on disk that no hook config references:\n\n```bash\nfor f in ~\u002F.claude\u002Fhooks\u002F*; do\n  name=$(basename \"$f\")\n  grep -rq \"$name\" ~\u002F.claude\u002Fsettings.json ~\u002F.claude\u002Fsettings.local.json 2>\u002Fdev\u002Fnull \\\n    || echo \"ORPHAN: $f\"\ndone\n```\n\nRedundant permission entries (channel 4) — duplicates, and specific grants shadowed by a wildcard:\n\n```bash\njq -r '.permissions.allow[]' ~\u002F.claude\u002Fsettings.local.json | sort | uniq -d\nif jq -e '.permissions.allow | index(\"Bash(*)\")' ~\u002F.claude\u002Fsettings.local.json >\u002Fdev\u002Fnull; then\n  jq -r '.permissions.allow[]' ~\u002F.claude\u002Fsettings.local.json \\\n    | grep '^Bash(' | grep -vF 'Bash(*)'\nfi\n```\n\nLargest stale caches (channel 8) — `du -k` instead of GNU-only `find -printf`, so it works on macOS\u002FBSD too:\n\n```bash\nfind ~\u002F.claude\u002Ffile-history ~\u002F.claude\u002Fshell-snapshots -type f -mtime +30 \\\n  -exec du -k {} + 2>\u002Fdev\u002Fnull | sort -rn | head -20\n```\n\nSoft-delete with undo path (capture the date once so the log can't disagree with the directory):\n\n```bash\ngc_date=$(date +%Y-%m-%d)\nmkdir -p ~\u002F.claude\u002F_gc_trash\u002F$gc_date\nmv ~\u002F.claude\u002Fskills\u002Fdead-skill ~\u002F.claude\u002F_gc_trash\u002F$gc_date\u002F\necho \"$(date -Iseconds) moved skills\u002Fdead-skill -> _gc_trash\u002F$gc_date\u002F (undo: mv back)\" >> ~\u002F.claude\u002Fgc_log.md\n```\n\nRemoving a confirmed-redundant permission entry (JSON has no comments — back up, log, then edit):\n\n```bash\ncp ~\u002F.claude\u002Fsettings.local.json ~\u002F.claude\u002Fsettings.local.json.bak\necho \"$(date -Iseconds) removed permission entry: Bash(git push) (undo: restore from .bak or re-add)\" >> ~\u002F.claude\u002Fgc_log.md\njq '.permissions.allow -= [\"Bash(git push)\"]' ~\u002F.claude\u002Fsettings.local.json.bak \\\n  > ~\u002F.claude\u002Fsettings.local.json\n```\n\n## Anti-Patterns\n\n- **Bulk approval.** Asking \"delete all 15? [y\u002Fn]\" defeats the design. One item, one decision.\n- **Hard-deleting on first pass.** If there's no `_gc_trash\u002F` copy or `.disabled` rename, you did it wrong.\n- **Treating \"old\" as \"dead\".** A skill untouched for 60 days may be seasonal (tax season, quarterly reviews). Age is a signal, not a verdict — that's why a human confirms.\n- **Cleaning memory by truncation.** Merging two contradicting memory files requires reading both and keeping the newer truth, not deleting the longer one.\n- **Touching anything outside `~\u002F.claude`** (or the project's `.claude\u002F`). Config GC never wanders into source trees.\n\n## Best Practices\n\n- Run after big additions, not just on a calendar: installing a 50-skill pack is exactly when overlap with existing skills appears.\n- When two skills overlap, prefer disabling the one with the weaker trigger description — it's the one that was probably never firing anyway.\n- Permission cleanup is the highest-value channel per minute spent: redundant allow-entries make security review harder.\n- Keep `gc_log.md` forever. It's tiny, and \"when did I disable that hook and why\" comes up more often than you'd think.\n\n## Related Skills\n\n- `skill-stocktake` — audits skill *quality*; config-gc audits skill *existence*. Run stocktake on what survives GC.\n- `workspace-surface-audit` — the additive counterpart: recommends what to install. config-gc is the subtractive half of the same lifecycle.\n- `configure-ecc` — after installing skills with it, run config-gc to reconcile overlaps with your pre-existing setup.\n- `continuous-learning` — produces the memory files this skill later audits.\n- `security-review` — pairs well with the permissions channel.\n",{"name":5,"metadata":24,"description":7},{"origin":25},"ECC","main","5d0a87185c0e172c3901e2a6e5916b6d544b8b94","skills\u002Fconfig-gc","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance","https:\u002F\u002Fgithub.com\u002FJunMystery","AI is the typist, I am the architect. Building efficient, secure, and user-centric products through disciplined AI prompt engineering.","2026-11-13 17:41:25",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add JunMystery\u002FAgent-Guidance --skill config-gc",{"tool":39,"command":40},"bunx","bunx skills add JunMystery\u002FAgent-Guidance --skill config-gc",{"tool":42,"command":43},"pnpm","pnpm dlx skills add JunMystery\u002FAgent-Guidance --skill config-gc","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fconfig-gc\u002FSKILL.md",true,[],[48,54,60,66,72,78],{"skillNo":49,"name":50,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M79QR1E309VJ6CTQKB32AWWA","zero-trust-architecture-cheat-sheet","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fzero-trust-architecture-cheat-sheet\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M79QQV7GHYR02DYGN41S23WN","xss-filter-evasion-cheat-sheet","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fxss-filter-evasion-cheat-sheet\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M79QQNSN7YB8XB61FFSCS3WH","xs-leaks-cheat-sheet","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fxs-leaks-cheat-sheet\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M79QQHZZJKFDY9132G5A65C2","xpath-injection-prevention-cheat-sheet","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fxpath-injection-prevention-cheat-sheet\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M79QQB9ZSD3DJG8658EHQXR1","xml-security-cheat-sheet","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fxml-security-cheat-sheet\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M79QQ60WKV7PFMXWT9CW8PM6","xml-external-entity-prevention-cheat-sheet","https:\u002F\u002Fgithub.com\u002FJunMystery\u002FAgent-Guidance\u002Fblob\u002Fmain\u002Fskills\u002Fxml-external-entity-prevention-cheat-sheet\u002FSKILL.md",[],[]]