[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EJZE55SX2QAPV8VTTPCYR3":3,"skill-same:01M7EJZE55SX2QAPV8VTTPCYR3":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":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M7EJZE55SX2QAPV8VTTPCYR3","happier-commit-worktree",null,"Reconnoiter, classify, validate, group, and commit a large or continuously changing Happier worktree as coherent, human-understandable commits while preserving concurrent work and excluding temporary, generated, QA, evidence, build, and other unwanted artifacts. Use when the user asks to commit many existing uncommitted changes, continue a long-running commit campaign, explain what remains, recover that campaign after compaction or interruption, or safely process newly landed changes in a shared dirty checkout. Do not use for an ordinary single-purpose commit whose scope is already known.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: happier-commit-worktree\ndescription: Reconnoiter, classify, validate, group, and commit a large or continuously changing Happier worktree as coherent, human-understandable commits while preserving concurrent work and excluding temporary, generated, QA, evidence, build, and other unwanted artifacts. Use when the user asks to commit many existing uncommitted changes, continue a long-running commit campaign, explain what remains, recover that campaign after compaction or interruption, or safely process newly landed changes in a shared dirty checkout. Do not use for an ordinary single-purpose commit whose scope is already known.\n---\n\n# Happier Commit Worktree\n\nTurn an existing moving worktree into a sequence of reviewable commits without treating file count, directory boundaries, or the current index as truth. Preserve all bytes on disk, prove what each commit contains, and leave uncertain material uncommitted with a specific reason.\n\nA large-worktree request is a campaign, not a request for one commit or one analysis wave. Continue recon, packet preparation, committing, and residual classification until every current path is committed or has an evidence-backed exclusion or blocker. Producing a few commits while commit-ready paths remain is incomplete execution unless the user explicitly pauses the campaign.\n\n## 1. Establish authority and safety\n\nRequire an explicit user request to commit. Analysis alone does not authorize staging or commits.\n\nApply these invariants throughout:\n\n- Never run `git reset`, `git restore`, `git clean`, `git checkout`, `git switch`, or an equivalent destructive operation.\n- Never overwrite, delete, or normalize unrelated work merely to make status clean.\n- Never trust or wholesale-commit the shared index. Stage every commit from explicit paths or hunks in a fresh private index.\n- Treat the checkout as live. A path can change before, during, or after a commit.\n- Serialize HEAD mutations even when reconnaissance and validation run in parallel.\n- Use Conventional Commit subjects and explanatory bodies.\n- Commit only changes whose intent, ownership, and suitability are understood.\n- Use the checkout's existing current-user `git config user.name` and `git config user.email` for every ordinary commit. Verify both before the first commit; never rewrite them to a bot, PR author, issue author, or other contributor, and stop rather than inventing a missing identity.\n- Evaluate `Co-authored-by:` attribution per packet. Add a verified contributor only when that packet materially incorporates their code, patch, design, causal diagnosis, decisive reproduction, or substantially adopted fix direction; never infer attribution from PR\u002Fissue authorship or participation alone.\n\nRead [private-index-protocol.md](references\u002Fprivate-index-protocol.md) before the first commit in a campaign. Read [recovery-and-audit.md](references\u002Frecovery-and-audit.md) whenever the index is unusual, a process was interrupted, HEAD moved, a lock appeared, or the user asks whether everything was preserved.\n\n## 2. Snapshot the current basis\n\nRecord compact observed evidence before grouping:\n\n```bash\ngit rev-parse HEAD\ngit status --short\ngit diff --cached --name-status\ngit diff --stat\ngit ls-files --others --exclude-standard\n```\n\nDo not equate a clean shared index with a clean worktree. Do not equate an untracked path with source code. If the shared index is non-empty, inspect it before proceeding; never clear inherited staging by assumption.\n\nBuild a resumable inventory grouped by package, domain, change kind, and likely provenance. Snapshot at wave boundaries and refresh paths that overlap a completed commit or changed during analysis; do not repeat repository-wide recon after every commit when independent inventory remains valid.\n\n## 3. Run reconnaissance before committing\n\nUse maximum useful parallelism where it shortens the critical path. Give each lane an exact, preferably disjoint path inventory and require:\n\n- candidate groups with exact paths and any hunk-level splits;\n- the user or system behavior each group establishes;\n- canonical owner, coupled tests, fixtures, schemas, docs, and generated contracts;\n- suspected artifacts or uncertain paths and the evidence for exclusion;\n- dependency ordering and validation commands;\n- a proposed Conventional Commit subject and explanatory body.\n\nThe lane must account for every assigned path as part of a commit-ready packet, an evidence-backed exclusion, or an unresolved item with the exact missing fact. It does not finish after finding representative groups or the first few commits. Sampling is not successful reconnaissance.\n\nReconnaissance lanes do not stage or mutate Git unless explicitly assigned commit authority. A single orchestrator should normally perform HEAD updates. If multiple agents must commit, each uses a private index and compare-and-swap update from its observed parent; a stale agent rebuilds from the new HEAD rather than forcing or replaying blindly.\n\nFollow the active subagent context policy. Default to no inherited transcript and provide a self-contained brief with exact scope, evidence, paths, checks, output, and stop conditions. Do not let lanes create ad hoc review files or use path \"custody\" as a substitute for checking current bytes and actual hunk collisions.\n\nRead [campaign-throughput.md](references\u002Fcampaign-throughput.md) before orchestrating a large or continuing campaign. It defines rolling waves, confidence lanes, packet queues, parallel topology, validation reuse, progress gates, compaction anchors, and valid stopping conditions. Read [grouping-and-messages.md](references\u002Fgrouping-and-messages.md) for the classification rubric, artifact policy, commit sizing, and message standard.\n\n## 4. Operate a rolling commit campaign\n\nKeep the serial commit authority supplied with prepared work:\n\n1. classify current paths into green, yellow, and red confidence lanes;\n2. prepare several coherent packets ahead, commonly 5-15 when scope permits;\n3. validate independent packets concurrently and drain ready packets through serial HEAD updates;\n4. re-snapshot after the wave and refresh only overlapping, newly landed, or invalidated paths;\n5. repeat while any commit-ready or safely investigable source work remains.\n\nGreen packets proceed immediately. Yellow investigation and red exclusions must not idle unrelated green work. Maintain a compact in-memory queue containing each packet's intent, owner, exact paths\u002Fhunks, dependencies, message, validation, and confidence. If many valid paths remain but the ready queue is empty, reconnaissance has failed and must resume rather than ending the campaign.\n\nParallelize reconnaissance, history\u002Fprovenance checks, message preparation, and independent validation. Keep final packet adjudication, private-index creation, CAS HEAD updates, and shared-index synchronization under one serial authority by default. Parallel commit writers are exceptional: private indexes isolate staging but not history, so CAS retries and overlap recovery can cost more than they save.\n\n## 5. Form coherent change packets\n\nGroup by one reviewable intent, invariant, migration, or user outcome, not by arbitrary path count. Include the complete slice needed to understand and verify that intent:\n\n- canonical implementation and its directly coupled adapters;\n- tests that define the behavior;\n- required fixtures, types, schemas, migrations, and compatibility handling;\n- consumed wiring and narrowly coupled documentation;\n- tracked generated output only when the repository contract requires it.\n\nPrefer useful batches, commonly 5-50 files and sometimes larger for uniform mechanical migrations, provider matrices, icon replacements, snapshots, or generated-contract updates. File count is a throughput heuristic, never permission to mix unrelated work. Avoid one-file commits when nearby changes complete the same idea, but keep a truly self-contained one-file correction separate.\n\nSplit a file by hunk when its changes serve different intents. Do not force an entire mixed file into the first convenient group.\n\nOrder packets by dependency: contracts and shared owners before consumers, implementation with defining tests, migrations before cleanup, and mechanical follow-ups after behavior is established.\n\n## 6. Reject dumping grounds and unwanted material\n\nBefore staging an unfamiliar path, determine what produced it and whether it belongs in source control. Use current bytes, repository references, ignore rules, tracked history, build scripts, test harnesses, and timestamps as evidence. Typical exclusions include:\n\n- build products, caches, compiler output, vendored downloads, archives, and oversized binaries;\n- `.tmp*`, `.vite-node`, probes, logs, screenshots, recordings, coverage, evidence, and QA captures;\n- local credentials, machine state, private diagnostics, and uploads;\n- semantic no-ops, accidental formatting, obsolete scratch code, and speculative dormant mechanisms;\n- tests or helpers whose production owner was removed unless the deletion itself is intentional and coherent.\n\nDo not delete uncertain paths as part of committing. Leave them uncommitted and report why. Add a narrow ignore rule when the producer is legitimate, recurrence is likely, and the path class is never source material. Do not hide a tracked source path or broad directory merely to make status disappear.\n\n## 7. Validate packets without serializing avoidable work\n\nInspect the exact private-index diff before creating a commit:\n\n```bash\nGIT_INDEX_FILE=\"$idx\" git diff --cached --stat\nGIT_INDEX_FILE=\"$idx\" git diff --cached --check\nGIT_INDEX_FILE=\"$idx\" git diff --cached --name-status\nGIT_INDEX_FILE=\"$idx\" git diff --cached\n```\n\nRun the narrowest deciding tests appropriate to each packet. Batch compatible package-wide typechecks, builds, and broader suites once per wave rather than repeating an identical expensive check after every commit. Record which packets a shared check covers and invalidate that evidence only when later bytes touch its deciding corridor. For pre-existing behavior changes, inspect whether tests and implementation agree rather than automatically changing whichever fails. Classify failures as a real defect, test drift, harness\u002Fenvironment failure, external-contract change, resource saturation, or unrelated concurrent failure.\n\n`git commit-tree` does not run ordinary `pre-commit`, `prepare-commit-msg`, `commit-msg`, or `post-commit` hooks. Before the campaign's first commit, inspect repository hook policy and run the required hook-equivalent checks explicitly for every applicable packet and message. Preserve required signing policy rather than silently creating unsigned commits.\n\nCheck for accidental secrets and objects that violate the remote's file-size policy before committing large or binary material. Never solve a source-control size failure by assuming Git LFS or committing a vendor\u002Fbuild tree without establishing that repository policy requires it.\n\nDo not claim a test or typecheck passed unless it ran. A coherent commit may proceed with a known unrelated failing check or unavailable saturated test environment only when the limitation is evidenced, unaffected, and disclosed. Do not repeatedly launch checks known to be infrastructure-blocked; continue independent packets and retry at a useful wave boundary.\n\n## 8. Commit from a private index\n\nFollow [private-index-protocol.md](references\u002Fprivate-index-protocol.md) exactly. The essential transaction is:\n\n1. capture the current `HEAD` as the intended parent;\n2. initialize a unique temporary index from that parent;\n3. stage only the packet's explicit current paths and selected hunks;\n4. inspect and validate the private staged diff;\n5. create a commit object with `git commit-tree`;\n6. advance `HEAD` with `git update-ref HEAD \u003Ccommit> \u003Cexpected-parent>`;\n7. synchronize only committed paths in the shared index to the new HEAD;\n8. verify the shared index and selected worktree paths.\n\nCompare-and-swap is mandatory. If HEAD moved, discard only the temporary index and rebuild the packet from the new HEAD and current worktree bytes. Never force the ref and never assume the previously built tree can simply be attached to a different parent.\n\nThis transaction never rewrites the worktree. If new bytes land in a committed file after private staging, the committed snapshot becomes HEAD and the newer bytes remain visible as an uncommitted modification. That is the required behavior.\n\n## 9. Reconcile after every commit and measure the wave\n\nImmediately verify:\n\n```bash\ngit show --stat --oneline --decorate -1\ngit diff --cached --name-status\ngit status --short -- \u003Ccommitted-paths...>\n```\n\nAn `M` after the commit can be correct: compare HEAD, index, and worktree to determine whether later bytes remain. A staged deletion after a successful commit is usually an index-synchronization defect; explicitly remove the deleted path from the shared index as documented in the protocol.\n\nMaintain a compact campaign ledger in the conversation or an already-approved tracking document, not a new ad hoc report file. Track completed commit ids, domain coverage, validation, residual groups, exclusions, and blockers. This is the anchor after compaction or interruption; always inspect live Git state before trusting it.\n\nAt each wave boundary record the starting and remaining path counts, paths consumed, commits created, ready queue depth, exclusions, unresolved count, current HEAD, next prepared packets, and validation constraints. Progress is reduced unresolved work, not merely commit count. Preserve this anchor in every compaction or continuation handoff so the next agent resumes instead of restarting recon.\n\n## 10. Close only after an exhaustive residual audit\n\nWhen no clear candidate group remains:\n\n1. re-snapshot HEAD, shared index, tracked modifications, deletions, and untracked paths;\n2. classify every residual path as newly landed source work, intentional artifact\u002Fexclusion, uncertain material, or an index anomaly;\n3. verify no candidate group was lost between reconnaissance and commit;\n4. inspect recent commits and aggregate their file\u002Fchange statistics when requested;\n5. report what was committed, checks run, failures or skipped checks, and every intentional residual category.\n\n\"Uncertain\" is not a convenient stopping label. Before using it, inspect the current diff, references\u002Fcallers, relevant history, owner\u002Ftest relationship, and provenance evidence appropriate to the path. State the exact decision that remains unresolved and the observation that would decide it. Continue all independent work that cannot prejudge that decision.\n\nDo not stop because one wave completed, one domain is exhausted, an agent returned partial results, a preferred test environment is saturated, or some red paths remain. Stop only when every current path is committed or classified, no commit-ready packet remains, and safely obtainable evidence cannot resolve the remaining blockers; or when the user explicitly pauses.\n\nSay \"everything appropriate is committed\" only when every current residual path has an observed reason not to commit. Never shorten that to \"everything is committed\" when artifacts, uncertainty, later changes, or blockers remain.\n\nBefore handoff, run `.agents\u002Fskills\u002Fattack-conclusion`: challenge grouping coherence, missing siblings, artifact provenance, index cleanliness, concurrent-byte preservation, dependency ordering, and the possibility that a passing check did not exercise the changed contract.\n","# Happier Commit Worktree\n\nTurn an existing moving worktree into a sequence of reviewable commits without treating file count, directory boundaries, or the current index as truth. Preserve all bytes on disk, prove what each commit contains, and leave uncertain material uncommitted with a specific reason.\n\nA large-worktree request is a campaign, not a request for one commit or one analysis wave. Continue recon, packet preparation, committing, and residual classification until every current path is committed or has an evidence-backed exclusion or blocker. Producing a few commits while commit-ready paths remain is incomplete execution unless the user explicitly pauses the campaign.\n\n## 1. Establish authority and safety\n\nRequire an explicit user request to commit. Analysis alone does not authorize staging or commits.\n\nApply these invariants throughout:\n\n- Never run `git reset`, `git restore`, `git clean`, `git checkout`, `git switch`, or an equivalent destructive operation.\n- Never overwrite, delete, or normalize unrelated work merely to make status clean.\n- Never trust or wholesale-commit the shared index. Stage every commit from explicit paths or hunks in a fresh private index.\n- Treat the checkout as live. A path can change before, during, or after a commit.\n- Serialize HEAD mutations even when reconnaissance and validation run in parallel.\n- Use Conventional Commit subjects and explanatory bodies.\n- Commit only changes whose intent, ownership, and suitability are understood.\n- Use the checkout's existing current-user `git config user.name` and `git config user.email` for every ordinary commit. Verify both before the first commit; never rewrite them to a bot, PR author, issue author, or other contributor, and stop rather than inventing a missing identity.\n- Evaluate `Co-authored-by:` attribution per packet. Add a verified contributor only when that packet materially incorporates their code, patch, design, causal diagnosis, decisive reproduction, or substantially adopted fix direction; never infer attribution from PR\u002Fissue authorship or participation alone.\n\nRead [private-index-protocol.md](references\u002Fprivate-index-protocol.md) before the first commit in a campaign. Read [recovery-and-audit.md](references\u002Frecovery-and-audit.md) whenever the index is unusual, a process was interrupted, HEAD moved, a lock appeared, or the user asks whether everything was preserved.\n\n## 2. Snapshot the current basis\n\nRecord compact observed evidence before grouping:\n\n```bash\ngit rev-parse HEAD\ngit status --short\ngit diff --cached --name-status\ngit diff --stat\ngit ls-files --others --exclude-standard\n```\n\nDo not equate a clean shared index with a clean worktree. Do not equate an untracked path with source code. If the shared index is non-empty, inspect it before proceeding; never clear inherited staging by assumption.\n\nBuild a resumable inventory grouped by package, domain, change kind, and likely provenance. Snapshot at wave boundaries and refresh paths that overlap a completed commit or changed during analysis; do not repeat repository-wide recon after every commit when independent inventory remains valid.\n\n## 3. Run reconnaissance before committing\n\nUse maximum useful parallelism where it shortens the critical path. Give each lane an exact, preferably disjoint path inventory and require:\n\n- candidate groups with exact paths and any hunk-level splits;\n- the user or system behavior each group establishes;\n- canonical owner, coupled tests, fixtures, schemas, docs, and generated contracts;\n- suspected artifacts or uncertain paths and the evidence for exclusion;\n- dependency ordering and validation commands;\n- a proposed Conventional Commit subject and explanatory body.\n\nThe lane must account for every assigned path as part of a commit-ready packet, an evidence-backed exclusion, or an unresolved item with the exact missing fact. It does not finish after finding representative groups or the first few commits. Sampling is not successful reconnaissance.\n\nReconnaissance lanes do not stage or mutate Git unless explicitly assigned commit authority. A single orchestrator should normally perform HEAD updates. If multiple agents must commit, each uses a private index and compare-and-swap update from its observed parent; a stale agent rebuilds from the new HEAD rather than forcing or replaying blindly.\n\nFollow the active subagent context policy. Default to no inherited transcript and provide a self-contained brief with exact scope, evidence, paths, checks, output, and stop conditions. Do not let lanes create ad hoc review files or use path \"custody\" as a substitute for checking current bytes and actual hunk collisions.\n\nRead [campaign-throughput.md](references\u002Fcampaign-throughput.md) before orchestrating a large or continuing campaign. It defines rolling waves, confidence lanes, packet queues, parallel topology, validation reuse, progress gates, compaction anchors, and valid stopping conditions. Read [grouping-and-messages.md](references\u002Fgrouping-and-messages.md) for the classification rubric, artifact policy, commit sizing, and message standard.\n\n## 4. Operate a rolling commit campaign\n\nKeep the serial commit authority supplied with prepared work:\n\n1. classify current paths into green, yellow, and red confidence lanes;\n2. prepare several coherent packets ahead, commonly 5-15 when scope permits;\n3. validate independent packets concurrently and drain ready packets through serial HEAD updates;\n4. re-snapshot after the wave and refresh only overlapping, newly landed, or invalidated paths;\n5. repeat while any commit-ready or safely investigable source work remains.\n\nGreen packets proceed immediately. Yellow investigation and red exclusions must not idle unrelated green work. Maintain a compact in-memory queue containing each packet's intent, owner, exact paths\u002Fhunks, dependencies, message, validation, and confidence. If many valid paths remain but the ready queue is empty, reconnaissance has failed and must resume rather than ending the campaign.\n\nParallelize reconnaissance, history\u002Fprovenance checks, message preparation, and independent validation. Keep final packet adjudication, private-index creation, CAS HEAD updates, and shared-index synchronization under one serial authority by default. Parallel commit writers are exceptional: private indexes isolate staging but not history, so CAS retries and overlap recovery can cost more than they save.\n\n## 5. Form coherent change packets\n\nGroup by one reviewable intent, invariant, migration, or user outcome, not by arbitrary path count. Include the complete slice needed to understand and verify that intent:\n\n- canonical implementation and its directly coupled adapters;\n- tests that define the behavior;\n- required fixtures, types, schemas, migrations, and compatibility handling;\n- consumed wiring and narrowly coupled documentation;\n- tracked generated output only when the repository contract requires it.\n\nPrefer useful batches, commonly 5-50 files and sometimes larger for uniform mechanical migrations, provider matrices, icon replacements, snapshots, or generated-contract updates. File count is a throughput heuristic, never permission to mix unrelated work. Avoid one-file commits when nearby changes complete the same idea, but keep a truly self-contained one-file correction separate.\n\nSplit a file by hunk when its changes serve different intents. Do not force an entire mixed file into the first convenient group.\n\nOrder packets by dependency: contracts and shared owners before consumers, implementation with defining tests, migrations before cleanup, and mechanical follow-ups after behavior is established.\n\n## 6. Reject dumping grounds and unwanted material\n\nBefore staging an unfamiliar path, determine what produced it and whether it belongs in source control. Use current bytes, repository references, ignore rules, tracked history, build scripts, test harnesses, and timestamps as evidence. Typical exclusions include:\n\n- build products, caches, compiler output, vendored downloads, archives, and oversized binaries;\n- `.tmp*`, `.vite-node`, probes, logs, screenshots, recordings, coverage, evidence, and QA captures;\n- local credentials, machine state, private diagnostics, and uploads;\n- semantic no-ops, accidental formatting, obsolete scratch code, and speculative dormant mechanisms;\n- tests or helpers whose production owner was removed unless the deletion itself is intentional and coherent.\n\nDo not delete uncertain paths as part of committing. Leave them uncommitted and report why. Add a narrow ignore rule when the producer is legitimate, recurrence is likely, and the path class is never source material. Do not hide a tracked source path or broad directory merely to make status disappear.\n\n## 7. Validate packets without serializing avoidable work\n\nInspect the exact private-index diff before creating a commit:\n\n```bash\nGIT_INDEX_FILE=\"$idx\" git diff --cached --stat\nGIT_INDEX_FILE=\"$idx\" git diff --cached --check\nGIT_INDEX_FILE=\"$idx\" git diff --cached --name-status\nGIT_INDEX_FILE=\"$idx\" git diff --cached\n```\n\nRun the narrowest deciding tests appropriate to each packet. Batch compatible package-wide typechecks, builds, and broader suites once per wave rather than repeating an identical expensive check after every commit. Record which packets a shared check covers and invalidate that evidence only when later bytes touch its deciding corridor. For pre-existing behavior changes, inspect whether tests and implementation agree rather than automatically changing whichever fails. Classify failures as a real defect, test drift, harness\u002Fenvironment failure, external-contract change, resource saturation, or unrelated concurrent failure.\n\n`git commit-tree` does not run ordinary `pre-commit`, `prepare-commit-msg`, `commit-msg`, or `post-commit` hooks. Before the campaign's first commit, inspect repository hook policy and run the required hook-equivalent checks explicitly for every applicable packet and message. Preserve required signing policy rather than silently creating unsigned commits.\n\nCheck for accidental secrets and objects that violate the remote's file-size policy before committing large or binary material. Never solve a source-control size failure by assuming Git LFS or committing a vendor\u002Fbuild tree without establishing that repository policy requires it.\n\nDo not claim a test or typecheck passed unless it ran. A coherent commit may proceed with a known unrelated failing check or unavailable saturated test environment only when the limitation is evidenced, unaffected, and disclosed. Do not repeatedly launch checks known to be infrastructure-blocked; continue independent packets and retry at a useful wave boundary.\n\n## 8. Commit from a private index\n\nFollow [private-index-protocol.md](references\u002Fprivate-index-protocol.md) exactly. The essential transaction is:\n\n1. capture the current `HEAD` as the intended parent;\n2. initialize a unique temporary index from that parent;\n3. stage only the packet's explicit current paths and selected hunks;\n4. inspect and validate the private staged diff;\n5. create a commit object with `git commit-tree`;\n6. advance `HEAD` with `git update-ref HEAD \u003Ccommit> \u003Cexpected-parent>`;\n7. synchronize only committed paths in the shared index to the new HEAD;\n8. verify the shared index and selected worktree paths.\n\nCompare-and-swap is mandatory. If HEAD moved, discard only the temporary index and rebuild the packet from the new HEAD and current worktree bytes. Never force the ref and never assume the previously built tree can simply be attached to a different parent.\n\nThis transaction never rewrites the worktree. If new bytes land in a committed file after private staging, the committed snapshot becomes HEAD and the newer bytes remain visible as an uncommitted modification. That is the required behavior.\n\n## 9. Reconcile after every commit and measure the wave\n\nImmediately verify:\n\n```bash\ngit show --stat --oneline --decorate -1\ngit diff --cached --name-status\ngit status --short -- \u003Ccommitted-paths...>\n```\n\nAn `M` after the commit can be correct: compare HEAD, index, and worktree to determine whether later bytes remain. A staged deletion after a successful commit is usually an index-synchronization defect; explicitly remove the deleted path from the shared index as documented in the protocol.\n\nMaintain a compact campaign ledger in the conversation or an already-approved tracking document, not a new ad hoc report file. Track completed commit ids, domain coverage, validation, residual groups, exclusions, and blockers. This is the anchor after compaction or interruption; always inspect live Git state before trusting it.\n\nAt each wave boundary record the starting and remaining path counts, paths consumed, commits created, ready queue depth, exclusions, unresolved count, current HEAD, next prepared packets, and validation constraints. Progress is reduced unresolved work, not merely commit count. Preserve this anchor in every compaction or continuation handoff so the next agent resumes instead of restarting recon.\n\n## 10. Close only after an exhaustive residual audit\n\nWhen no clear candidate group remains:\n\n1. re-snapshot HEAD, shared index, tracked modifications, deletions, and untracked paths;\n2. classify every residual path as newly landed source work, intentional artifact\u002Fexclusion, uncertain material, or an index anomaly;\n3. verify no candidate group was lost between reconnaissance and commit;\n4. inspect recent commits and aggregate their file\u002Fchange statistics when requested;\n5. report what was committed, checks run, failures or skipped checks, and every intentional residual category.\n\n\"Uncertain\" is not a convenient stopping label. Before using it, inspect the current diff, references\u002Fcallers, relevant history, owner\u002Ftest relationship, and provenance evidence appropriate to the path. State the exact decision that remains unresolved and the observation that would decide it. Continue all independent work that cannot prejudge that decision.\n\nDo not stop because one wave completed, one domain is exhausted, an agent returned partial results, a preferred test environment is saturated, or some red paths remain. Stop only when every current path is committed or classified, no commit-ready packet remains, and safely obtainable evidence cannot resolve the remaining blockers; or when the user explicitly pauses.\n\nSay \"everything appropriate is committed\" only when every current residual path has an observed reason not to commit. Never shorten that to \"everything is committed\" when artifacts, uncertainty, later changes, or blockers remain.\n\nBefore handoff, run `.agents\u002Fskills\u002Fattack-conclusion`: challenge grouping coherence, missing siblings, artifact provenance, index cleanliness, concurrent-byte preservation, dependency ordering, and the possibility that a passing check did not exercise the changed contract.\n",{"name":5,"description":7},"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","04-安全渗透\u002Fhappier-commit-worktree",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill happier-commit-worktree",{"tool":38,"command":39},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill happier-commit-worktree",{"tool":41,"command":42},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill happier-commit-worktree","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002FSKILL.md",true,[46,52,56,60,64],{"path":47,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":51},"agents\u002Fopenai.yaml","REFERENCE","参考文档",334,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002Fagents\u002Fopenai.yaml",{"path":53,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":54,"sourceHtmlUrl":55},"references\u002Fcampaign-throughput.md",8981,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002Freferences\u002Fcampaign-throughput.md",{"path":57,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Fgrouping-and-messages.md",9288,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002Freferences\u002Fgrouping-and-messages.md",{"path":61,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Fprivate-index-protocol.md",8250,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002Freferences\u002Fprivate-index-protocol.md",{"path":65,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":66,"sourceHtmlUrl":67},"references\u002Frecovery-and-audit.md",6383,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fhappier-commit-worktree\u002Freferences\u002Frecovery-and-audit.md",[69,77,82,89,96,103],{"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":76},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":78,"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":79,"categories":80,"keywords":81,"publishedAt":76},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"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":76},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":90,"name":91,"nameZh":6,"description":92,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":76},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":97,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"publishedAt":76},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":104,"name":105,"nameZh":6,"description":106,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":107,"categories":108,"keywords":109,"publishedAt":76},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]