[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AX2TKVME5S7DQ37JDFGSV5":3,"skill-same:01M7AX2TKVME5S7DQ37JDFGSV5":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":29,"updatedAt":30,"install":31,"resources":44},"01M7AX2TKVME5S7DQ37JDFGSV5","master",null,"Lead mid to large work by interviewing the user, then running a light track (up to 3 independent units built by members) or a heavy track (contracts, leads), debating open design questions and always reviewing across models. Normal loading starts the play for the whole session; for a single change do not activate it; do it yourself, commit only your own paths, then spawn one reviewer on another model (worktree=true, base_branch=\u003Cyour SHA>, yolo, review format) with a read-only brief naming the original base, your head SHA, the ACs and the checks. Push only after approve.","mana-byte\u002Ftheater","mana-byte","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F93316844?v=4","01M7AX1XSX6E0DJHN1SA4MEAJ3","01M7AX1XTJYZBKJH863Y7N1D61",1,"MIT","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fmcp_plugins\u002Fbuiltin\u002Fbeads\u002Fskills\u002Fhierarchy\u002Fmaster\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: master\ndescription: \"Lead mid to large work by interviewing the user, then running a light track (up to 3 independent units built by members) or a heavy track (contracts, leads), debating open design questions and always reviewing across models. Normal loading starts the play for the whole session; for a single change do not activate it; do it yourself, commit only your own paths, then spawn one reviewer on another model (worktree=true, base_branch=\u003Cyour SHA>, yolo, review format) with a read-only brief naming the original base, your head SHA, the ACs and the checks. Push only after approve.\"\n---\n\n# Master\n\nUse this play's Beads board for explicit task tracking. Create a root epic with the goal and\nacceptance criteria, then milestones and separate implementation and review tasks. Give each\ndelegate its task ID and expected outcome; make review tasks depend on implementation tasks,\nand milestones depend on both. Link each finished job with `link_task_job`: the work job and its review\njob both to the same implementation task — work as `work`, a review as `review` naming the\nreviewed work job in `review_of`; the separate review task records only its own lifecycle and\noutcome. An identical link is a no-op duplicate: if a\ndispatch succeeded but the link call failed in transit, retry only `link_task_job`, never\nre-spawn or re-send; act on a semantic refusal rather than retrying it. A re-review links its\nexact new review handle with `review_of` set to the repair's exact work handle. Evidence\n(`source_basis`, `artifact_relation`) is not approval, closure or merge. `get_task(view=\"progress\")` then shows the linked work,\nreview and evidence without re-reading job facts in the detail view. Record the final outcome by\nclosing the root task with a truthful `outcome_status` from checks\u002Freview (`succeeded`, `failed` or\n`cancelled`; omission means unknown). Closing it does not end Theater membership.\n\n`load_skill(name=\"hierarchy\u002Fmaster\", preview=true)` reads this card without identity or play\nactivation; omitting `preview` activates the play for this session. Every agent you start must hold\none position, plus at most one role; a plain spawn is refused. If the work turns out to be\na single unit, tell the user, spawn one `hierarchy\u002Fmember`+`hierarchy\u002Fworker`, and review it as in\nstep 6.\n\nYou alone talk to the user. Positions you give: `hierarchy\u002Flead`, `hierarchy\u002Fmember`. Roles:\n`hierarchy\u002Fworker` (lead or member), `hierarchy\u002Freviewer` and `hierarchy\u002Fdebater` (member only).\n`whoami` lists the same choices. Children inherit nothing, so every brief is complete. Never\nhardcode a model: use `list_harnesses(include_models=true)`. Debaters differ from you and each other\nas far as models allow (another harness is the fallback; report any shortfall). A reviewer must run\na different model than every author, including whoever assembled the work (a lead, or you for the\nfinal review); if none exists, report it as a blocker. Pass `approval=\"yolo\"` for read-only roles\nand the user's chosen approval mode for every builder spawn.\n\nFollow the user's requested interaction: stay and `await_sessions` for work they expect you to\nfinish here; for authorized background work, give them all active handles and end the turn. Review\nand integration resume on the user's next message; MCP has no server-initiated turn. Judge every\nawait entry by `state` (a finished agent held by human presence returns `timeout` with state `done`;\ntell the user once and continue), and re-await pending work when staying. A child's `done` is not a\nverdict: check the repo.\n\n## 0. Setup\n\n- `whoami`, `list_harnesses(include_models=true)`, and the participant budget `B` from\n  `theater config --json` (`rails.budget`, checked against its daemon line; if they differ `B` is provisional, and a\n  `budget_exceeded` refusal's number wins).\n- `base = git rev-parse HEAD`. The user's uncommitted work is excluded from every stream; if it\n  matters, ask the user to commit first.\n- Create the integration worktree under the main repo root (the parent of\n  `git rev-parse --path-format=absolute --git-common-dir`):\n  `git worktree add -b \u003Cbranch> \u003Croot>\u002F.theater\u002Fintegration\u002F\u003Cname> \u003Cbase>`. Stubs, merges and\n  the full suite run only there. Never switch, stash, stage or commit in the user's checkout.\n- Use the scratchpad when a brief is large or separately started workers need a durable handoff;\n  record phase, contracts and handles there as needed. It is not proof of approval.\n\n## 1. Intake: features first\n\nSource, in order: a plan the user points you to, earlier explicit user decisions and authorization\n(their file, your own context, or a participant's `read_transcript`), else ask in batches of at most\n4 questions, each with a recommended default. Ask about features and outcomes, not code. Keep the\ngoal, scope, branch\u002Fdestination, approval mode, named retirement\u002Fcleanup authorizations, at most 12\nnumbered acceptance criteria (AC), and constraints. Reuse a still-applicable explicit yes across\nturns, compaction, checks and retries; ask only about missing, changed or materially out-of-scope\ndecisions. `yolo` authorizes execution per spawn, not publication or deletion. No approval mode has\na global default. Do not infer authorization from a scratchpad note.\n\n## 2. Track\n\nSurvey the code read-only and list the plan's units: for each, the paths it owns, its rough size,\nthe symbols it changes and the symbols it consumes. Default heavy; any unknown selects heavy. The\ntrack is **light** only if all hold: at most 3 units, each one worker's task (none would need a\nlead); pairwise-disjoint paths; no unit consumes a symbol another unit adds, removes or changes\n(else the consumer starts after the producer merges, from the integration branch head, or the track\nis heavy); ACs settled. Name the track and these facts in the validation. Debate (step 3) runs on\nheavy, and on either track whenever a design question is open: two or more viable options that\ndiffer in behaviour, compatibility, data risk, cost or scope.\n\n## 3. Debate\n\nLoad `theater-debate` for its evidence and consensus criteria only; this play replaces its one-peer\nsetup and exchange schedule with the rules below. You spawn two debaters (three only for\ncross-cutting work), each on a model different from yours and the others', with\n`skills=[\"hierarchy\u002Fmember\", \"hierarchy\u002Fdebater\"]`, `worktree=true`, `base_branch=\u003Cdebate base>`\n(`base`; after an upshift, the integration branch head), `response_format={\"$theater\": \"debate\"}`.\nGive all the same neutral brief (requirements, fixed commit, options if known, acceptance criteria,\npaths), never your conclusion. Run three exchanges: independent positions; the other side's evidence\nforwarded unedited, each rebuts or concedes; a verdict on your synthesis. Every claim must cite\ncode, a command, a source or a user requirement.\n\nVerify what you rely on: every decision in your report needs a citation you re-checked (`git show\n\u003Cdebate base>:\u003Cpath>` or the original source; at most 2 checks per decision plus both sides of each\ncontradiction, 15 in total). An unverifiable claim is never support: ask for a source or carry it as\nan open point. Then either ask the user more questions, technical only where the answer changes\nbehaviour, compatibility, data risk, cost or scope (two options and a recommendation, no\nimplementation detail), or write the report. At most 3 debate-then-ask cycles, then decide delegated\ntechnical choices and escalate the user's.\n\n## 4. Validation\n\nAfter a debate, a report of at most 400 words: goal; scope in and out; ACs; at most 7 decisions\n(choice, evidence, rejected option); risks and open points. Light with no debate: one message with\nthe brief (at most 150 words), the track facts, what light omits (leads, contracts) and the\ndecisions taken without debate (choice, rejected option, why not viable) or none, so the user can\nask for a debate. Either way it carries the execution line (track, approval mode, base commit and\nexcluded work, model availability, integration branch). Ask for validation only where prior explicit\nauthorization does not already cover that line and the decisions; record what covers it. Units in\nthe track facts are preliminary; recompute them after a debate. A material change needs a new\ndecision; a turn boundary, harmless check or retry does not.\n\n## 5. Build (after validation)\n\n**Heavy.** One owner per path. Shared hot files (registries, re-exports, migrations, lockfiles) stay\nyours, landed at integration. Write a contract for each cross-stream interface: provider, consumer,\nexact shape, the check that proves it; commit code-expressible ones as stubs on the integration\nbranch and record the commit as `integration SHA`. One lead per stream: spawn it with\n`skills=[\"hierarchy\u002Flead\"]` (add `\"hierarchy\u002Fworker\"` for a small stream), `worktree=true`,\n`base_branch=\u003Cintegration SHA>`, `response_format={\"$theater\": \"delegate\"}` and the validated\napproval mode. Its brief: owned and forbidden paths, contracts verbatim, ACs, stream-base SHA,\nchecks, participant allotment, the approval mode for its children, and applicable user\nauthorization. A contract change comes only from you and goes to every affected stream.\n\n**Light.** Initially the integration SHA is `base`; before a dependent unit starts, update it to the\nintegration branch head after its producer merges. Spawn each unit as `skills=[\"hierarchy\u002Fmember\",\n\"hierarchy\u002Fworker\"]` with the same spawn settings; the brief holds its owned and forbidden paths,\nACs, checks, applicable user authorization, the symbols it may change and the symbols it must not\nalter (those other units consume at the same time); a producer scheduled before its consumer may\nchange what the consumer waits for. Its reviewer gets the same lists.\n\nSize the tree for either track: live tree now (debaters included) + leads + workers + reviewers + 1\nfinal reviewer must be at most `B-1`. An agent keeps its slot until it dies; retiring a member frees\nit. Reuse idle members by `send` (never turn a debater or reviewer into a builder); otherwise retire\nfinished members or cut the unit count. For a simple handoff to a suitable existing peer, keep its\ncurrent play position and role: one `send` names the goal, owned and forbidden scope, expected\noutput, ACs, checks and applicable explicit authorization. Use `response_format` when acting on its\nresult; await that send's exact returned handle. Do not start a nested pair play just for the send,\ninfer permission from `addressable`, or silently delegate review. Record each stream-base. Leads and\nworkers escalate by ending their turn `blocked`.\n\n## 6. Review\n\nWhen a lead or worker reports `done`: audit `git diff --name-status -M \u003Cstream-base> \u003CSHA>` (both\nrename paths) against its owned paths. Then spawn one reviewer per stream or unit:\n`skills=[\"hierarchy\u002Fmember\", \"hierarchy\u002Freviewer\"]`, `task_kind=\"independent_review\"`,\n`worktree=true`, `after=\u003Cexact handle of its last done turn>`, no `base_branch`,\n`response_format={\"$theater\": \"review\"}`, on a model that meets the rule above. Findings go to the\nauthor as `send`; it repairs with new commits under still-applicable authorization. Re-review with\n`send(reviewer, after=\u003Cexact handle of the repair turn>)` and the review `response_format` (never a\nname: while a repair is queued it still selects the previous turn); the reviewer fast-forwards to\nthe new SHA. If it cannot, spawn a fresh\n`independent_review`. After 2 repair rounds, diagnose (`get_session_controls`, a narrow transcript\ntail) and escalate. It is reviewed only once a parsed `approve` covers the exact SHA the author\nreported.\n\n## 7. Integrate\n\nIn the integration worktree, merge each reviewed SHA (never a branch name) in dependency order,\nafter checking the reviewer's reported head equals the author's SHA. Run focused checks per merge,\nthen check cross-stream interfaces and run the full test suite. Skip the final review only if exactly\none unit or stream merged, as a fast-forward onto `base` (you added no stubs, fixes or merge\ncommits), and the integrated SHA is its reviewed SHA. Otherwise freeze the integrated SHA and\nspawn a final reviewer: an ordinary spawn,\nso no handoff card, `skills=[\"hierarchy\u002Fmember\", \"hierarchy\u002Freviewer\"]`, `worktree=true`,\n`base_branch=\u003Cintegrated SHA>`, `response_format={\"$theater\": \"review\"}`. Its brief names the\ncomparison base (`base`), the frozen head SHA and the ACs as requirements; completion needs a parsed\n`approve` covering that SHA. Any later edit voids it. On a red suite or `changes_requested`, send\neach finding to the owning author (repair, re-review, re-merge its exact SHA) or fix files you own,\nrerun the suite and spawn a fresh final review; escalate after 2 rounds.\n\nThen send one closing message: each AC as met, partial or unmet, suite counts, deviations and risks.\nFor merge into the user's branch, push, integration worktree\u002Fbranch removal, and termination of\nremaining leads (named, with ids), carry out each action already explicitly authorized for that\ndestination or those names; ask separately only for any uncovered action.\n\n## Upshift\n\nLight becomes heavy when a worker ends `blocked` on a missing interface, a review finds a\ndesign-level blocker, units pass 3, a hidden contract appears, or the user's scope grows. Stop\nsending the affected workers work (`interrupt_session` only a turn whose output is now wrong); if a\ndesign question opened, debate with the debate base set to the integration branch head; validate any\nnew or changed decision; write the contracts; add leads and debaters for the remaining work only.\nExisting members stay members. Recount the slots.\n\n## Retirement\n\n`whoami` `play.disposable` lists the positions you may retire with `terminate_session` without\nasking the user: here, members. Retire one once its result is collected and its commit is recorded,\nmerged or dropped: debaters after validation, a reviewer once its parsed `approve` covers the exact\nSHA and no re-review is due, a worker after its merge. Before each retirement, check the child's\n`play.position` in `list_participants` and pass its id, not a name. To abandon a running member,\n`interrupt_session` it, await its original job handle, then retire it. Any position not in\n`play.disposable` (leads, unpositioned children) still needs the user's named yes.\n\n## Cleanup\n\nRetire your remaining finished members. On the user's applicable named yes, terminate exactly those\nleads. Leads retire their own members; give the user the id of any survivor below a lead to kill\nwith `theater kill`. Report survivors. The play ends when you do.\n","# Master\n\nUse this play's Beads board for explicit task tracking. Create a root epic with the goal and\nacceptance criteria, then milestones and separate implementation and review tasks. Give each\ndelegate its task ID and expected outcome; make review tasks depend on implementation tasks,\nand milestones depend on both. Link each finished job with `link_task_job`: the work job and its review\njob both to the same implementation task — work as `work`, a review as `review` naming the\nreviewed work job in `review_of`; the separate review task records only its own lifecycle and\noutcome. An identical link is a no-op duplicate: if a\ndispatch succeeded but the link call failed in transit, retry only `link_task_job`, never\nre-spawn or re-send; act on a semantic refusal rather than retrying it. A re-review links its\nexact new review handle with `review_of` set to the repair's exact work handle. Evidence\n(`source_basis`, `artifact_relation`) is not approval, closure or merge. `get_task(view=\"progress\")` then shows the linked work,\nreview and evidence without re-reading job facts in the detail view. Record the final outcome by\nclosing the root task with a truthful `outcome_status` from checks\u002Freview (`succeeded`, `failed` or\n`cancelled`; omission means unknown). Closing it does not end Theater membership.\n\n`load_skill(name=\"hierarchy\u002Fmaster\", preview=true)` reads this card without identity or play\nactivation; omitting `preview` activates the play for this session. Every agent you start must hold\none position, plus at most one role; a plain spawn is refused. If the work turns out to be\na single unit, tell the user, spawn one `hierarchy\u002Fmember`+`hierarchy\u002Fworker`, and review it as in\nstep 6.\n\nYou alone talk to the user. Positions you give: `hierarchy\u002Flead`, `hierarchy\u002Fmember`. Roles:\n`hierarchy\u002Fworker` (lead or member), `hierarchy\u002Freviewer` and `hierarchy\u002Fdebater` (member only).\n`whoami` lists the same choices. Children inherit nothing, so every brief is complete. Never\nhardcode a model: use `list_harnesses(include_models=true)`. Debaters differ from you and each other\nas far as models allow (another harness is the fallback; report any shortfall). A reviewer must run\na different model than every author, including whoever assembled the work (a lead, or you for the\nfinal review); if none exists, report it as a blocker. Pass `approval=\"yolo\"` for read-only roles\nand the user's chosen approval mode for every builder spawn.\n\nFollow the user's requested interaction: stay and `await_sessions` for work they expect you to\nfinish here; for authorized background work, give them all active handles and end the turn. Review\nand integration resume on the user's next message; MCP has no server-initiated turn. Judge every\nawait entry by `state` (a finished agent held by human presence returns `timeout` with state `done`;\ntell the user once and continue), and re-await pending work when staying. A child's `done` is not a\nverdict: check the repo.\n\n## 0. Setup\n\n- `whoami`, `list_harnesses(include_models=true)`, and the participant budget `B` from\n  `theater config --json` (`rails.budget`, checked against its daemon line; if they differ `B` is provisional, and a\n  `budget_exceeded` refusal's number wins).\n- `base = git rev-parse HEAD`. The user's uncommitted work is excluded from every stream; if it\n  matters, ask the user to commit first.\n- Create the integration worktree under the main repo root (the parent of\n  `git rev-parse --path-format=absolute --git-common-dir`):\n  `git worktree add -b \u003Cbranch> \u003Croot>\u002F.theater\u002Fintegration\u002F\u003Cname> \u003Cbase>`. Stubs, merges and\n  the full suite run only there. Never switch, stash, stage or commit in the user's checkout.\n- Use the scratchpad when a brief is large or separately started workers need a durable handoff;\n  record phase, contracts and handles there as needed. It is not proof of approval.\n\n## 1. Intake: features first\n\nSource, in order: a plan the user points you to, earlier explicit user decisions and authorization\n(their file, your own context, or a participant's `read_transcript`), else ask in batches of at most\n4 questions, each with a recommended default. Ask about features and outcomes, not code. Keep the\ngoal, scope, branch\u002Fdestination, approval mode, named retirement\u002Fcleanup authorizations, at most 12\nnumbered acceptance criteria (AC), and constraints. Reuse a still-applicable explicit yes across\nturns, compaction, checks and retries; ask only about missing, changed or materially out-of-scope\ndecisions. `yolo` authorizes execution per spawn, not publication or deletion. No approval mode has\na global default. Do not infer authorization from a scratchpad note.\n\n## 2. Track\n\nSurvey the code read-only and list the plan's units: for each, the paths it owns, its rough size,\nthe symbols it changes and the symbols it consumes. Default heavy; any unknown selects heavy. The\ntrack is **light** only if all hold: at most 3 units, each one worker's task (none would need a\nlead); pairwise-disjoint paths; no unit consumes a symbol another unit adds, removes or changes\n(else the consumer starts after the producer merges, from the integration branch head, or the track\nis heavy); ACs settled. Name the track and these facts in the validation. Debate (step 3) runs on\nheavy, and on either track whenever a design question is open: two or more viable options that\ndiffer in behaviour, compatibility, data risk, cost or scope.\n\n## 3. Debate\n\nLoad `theater-debate` for its evidence and consensus criteria only; this play replaces its one-peer\nsetup and exchange schedule with the rules below. You spawn two debaters (three only for\ncross-cutting work), each on a model different from yours and the others', with\n`skills=[\"hierarchy\u002Fmember\", \"hierarchy\u002Fdebater\"]`, `worktree=true`, `base_branch=\u003Cdebate base>`\n(`base`; after an upshift, the integration branch head), `response_format={\"$theater\": \"debate\"}`.\nGive all the same neutral brief (requirements, fixed commit, options if known, acceptance criteria,\npaths), never your conclusion. Run three exchanges: independent positions; the other side's evidence\nforwarded unedited, each rebuts or concedes; a verdict on your synthesis. Every claim must cite\ncode, a command, a source or a user requirement.\n\nVerify what you rely on: every decision in your report needs a citation you re-checked (`git show\n\u003Cdebate base>:\u003Cpath>` or the original source; at most 2 checks per decision plus both sides of each\ncontradiction, 15 in total). An unverifiable claim is never support: ask for a source or carry it as\nan open point. Then either ask the user more questions, technical only where the answer changes\nbehaviour, compatibility, data risk, cost or scope (two options and a recommendation, no\nimplementation detail), or write the report. At most 3 debate-then-ask cycles, then decide delegated\ntechnical choices and escalate the user's.\n\n## 4. Validation\n\nAfter a debate, a report of at most 400 words: goal; scope in and out; ACs; at most 7 decisions\n(choice, evidence, rejected option); risks and open points. Light with no debate: one message with\nthe brief (at most 150 words), the track facts, what light omits (leads, contracts) and the\ndecisions taken without debate (choice, rejected option, why not viable) or none, so the user can\nask for a debate. Either way it carries the execution line (track, approval mode, base commit and\nexcluded work, model availability, integration branch). Ask for validation only where prior explicit\nauthorization does not already cover that line and the decisions; record what covers it. Units in\nthe track facts are preliminary; recompute them after a debate. A material change needs a new\ndecision; a turn boundary, harmless check or retry does not.\n\n## 5. Build (after validation)\n\n**Heavy.** One owner per path. Shared hot files (registries, re-exports, migrations, lockfiles) stay\nyours, landed at integration. Write a contract for each cross-stream interface: provider, consumer,\nexact shape, the check that proves it; commit code-expressible ones as stubs on the integration\nbranch and record the commit as `integration SHA`. One lead per stream: spawn it with\n`skills=[\"hierarchy\u002Flead\"]` (add `\"hierarchy\u002Fworker\"` for a small stream), `worktree=true`,\n`base_branch=\u003Cintegration SHA>`, `response_format={\"$theater\": \"delegate\"}` and the validated\napproval mode. Its brief: owned and forbidden paths, contracts verbatim, ACs, stream-base SHA,\nchecks, participant allotment, the approval mode for its children, and applicable user\nauthorization. A contract change comes only from you and goes to every affected stream.\n\n**Light.** Initially the integration SHA is `base`; before a dependent unit starts, update it to the\nintegration branch head after its producer merges. Spawn each unit as `skills=[\"hierarchy\u002Fmember\",\n\"hierarchy\u002Fworker\"]` with the same spawn settings; the brief holds its owned and forbidden paths,\nACs, checks, applicable user authorization, the symbols it may change and the symbols it must not\nalter (those other units consume at the same time); a producer scheduled before its consumer may\nchange what the consumer waits for. Its reviewer gets the same lists.\n\nSize the tree for either track: live tree now (debaters included) + leads + workers + reviewers + 1\nfinal reviewer must be at most `B-1`. An agent keeps its slot until it dies; retiring a member frees\nit. Reuse idle members by `send` (never turn a debater or reviewer into a builder); otherwise retire\nfinished members or cut the unit count. For a simple handoff to a suitable existing peer, keep its\ncurrent play position and role: one `send` names the goal, owned and forbidden scope, expected\noutput, ACs, checks and applicable explicit authorization. Use `response_format` when acting on its\nresult; await that send's exact returned handle. Do not start a nested pair play just for the send,\ninfer permission from `addressable`, or silently delegate review. Record each stream-base. Leads and\nworkers escalate by ending their turn `blocked`.\n\n## 6. Review\n\nWhen a lead or worker reports `done`: audit `git diff --name-status -M \u003Cstream-base> \u003CSHA>` (both\nrename paths) against its owned paths. Then spawn one reviewer per stream or unit:\n`skills=[\"hierarchy\u002Fmember\", \"hierarchy\u002Freviewer\"]`, `task_kind=\"independent_review\"`,\n`worktree=true`, `after=\u003Cexact handle of its last done turn>`, no `base_branch`,\n`response_format={\"$theater\": \"review\"}`, on a model that meets the rule above. Findings go to the\nauthor as `send`; it repairs with new commits under still-applicable authorization. Re-review with\n`send(reviewer, after=\u003Cexact handle of the repair turn>)` and the review `response_format` (never a\nname: while a repair is queued it still selects the previous turn); the reviewer fast-forwards to\nthe new SHA. If it cannot, spawn a fresh\n`independent_review`. After 2 repair rounds, diagnose (`get_session_controls`, a narrow transcript\ntail) and escalate. It is reviewed only once a parsed `approve` covers the exact SHA the author\nreported.\n\n## 7. Integrate\n\nIn the integration worktree, merge each reviewed SHA (never a branch name) in dependency order,\nafter checking the reviewer's reported head equals the author's SHA. Run focused checks per merge,\nthen check cross-stream interfaces and run the full test suite. Skip the final review only if exactly\none unit or stream merged, as a fast-forward onto `base` (you added no stubs, fixes or merge\ncommits), and the integrated SHA is its reviewed SHA. Otherwise freeze the integrated SHA and\nspawn a final reviewer: an ordinary spawn,\nso no handoff card, `skills=[\"hierarchy\u002Fmember\", \"hierarchy\u002Freviewer\"]`, `worktree=true`,\n`base_branch=\u003Cintegrated SHA>`, `response_format={\"$theater\": \"review\"}`. Its brief names the\ncomparison base (`base`), the frozen head SHA and the ACs as requirements; completion needs a parsed\n`approve` covering that SHA. Any later edit voids it. On a red suite or `changes_requested`, send\neach finding to the owning author (repair, re-review, re-merge its exact SHA) or fix files you own,\nrerun the suite and spawn a fresh final review; escalate after 2 rounds.\n\nThen send one closing message: each AC as met, partial or unmet, suite counts, deviations and risks.\nFor merge into the user's branch, push, integration worktree\u002Fbranch removal, and termination of\nremaining leads (named, with ids), carry out each action already explicitly authorized for that\ndestination or those names; ask separately only for any uncovered action.\n\n## Upshift\n\nLight becomes heavy when a worker ends `blocked` on a missing interface, a review finds a\ndesign-level blocker, units pass 3, a hidden contract appears, or the user's scope grows. Stop\nsending the affected workers work (`interrupt_session` only a turn whose output is now wrong); if a\ndesign question opened, debate with the debate base set to the integration branch head; validate any\nnew or changed decision; write the contracts; add leads and debaters for the remaining work only.\nExisting members stay members. Recount the slots.\n\n## Retirement\n\n`whoami` `play.disposable` lists the positions you may retire with `terminate_session` without\nasking the user: here, members. Retire one once its result is collected and its commit is recorded,\nmerged or dropped: debaters after validation, a reviewer once its parsed `approve` covers the exact\nSHA and no re-review is due, a worker after its merge. Before each retirement, check the child's\n`play.position` in `list_participants` and pass its id, not a name. To abandon a running member,\n`interrupt_session` it, await its original job handle, then retire it. Any position not in\n`play.disposable` (leads, unpositioned children) still needs the user's named yes.\n\n## Cleanup\n\nRetire your remaining finished members. On the user's applicable named yes, terminate exactly those\nleads. Leads retire their own members; give the user the id of any survivor below a lead to kill\nwith `theater kill`. Report survivors. The play ends when you do.\n",{"name":5,"description":7},"main","8ff9aa9ce8393ccad496ed8806324b7b4b6c2314","theater\u002Fmcp_plugins\u002Fbuiltin\u002Fbeads\u002Fskills\u002Fhierarchy\u002Fmaster",0,"https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater","https:\u002F\u002Fgithub.com\u002Fmana-byte","Always learning","2026-11-14 20:40:25",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add mana-byte\u002Ftheater --skill master",{"tool":37,"command":38},"bunx","bunx skills add mana-byte\u002Ftheater --skill master",{"tool":40,"command":41},"pnpm","pnpm dlx skills add mana-byte\u002Ftheater --skill master","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fmcp_plugins\u002Fbuiltin\u002Fbeads\u002Fskills\u002Fhierarchy\u002Fmaster\u002FSKILL.md",true,[],[46,53,60,67,74,81],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":18},"01M7AX3QM639GSJHJAAE6VCAT1","theater-recover-tmux","Recover Theater participants after a tmux server restart or kill-server using retained participant metadata, native session resume, and an explicitly chosen conservative recovery mode.","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fskills\u002Fbuiltin\u002Ftheater-recover-tmux\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":18},"01M7AX3M8JZ8FPRCP4F3CFWDN7","theater-delegate","Hand a scoped task to a suitable existing Theater peer, or start a builder and independent reviewer pair when the work needs one. Stay or run in the background as the user requested.","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fskills\u002Fbuiltin\u002Ftheater-delegate\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":18},"01M7AX3FDEF136J646DZWVW3K6","theater-debate","Debate a review, design, or architecture decision with one specified Theater participant when the user asks to debate with a harness, model, and reasoning effort. Use for evidence-backed disagreement, adversarial examination, and bounded convergence before making an important technical decision.","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fskills\u002Fbuiltin\u002Ftheater-debate\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":18},"01M7AX3AEFDJHJJMVYTPF79P61","theater-configure","Configure or personalize Theater interactively when the user wants to set up, review, or change their Theater settings — favourite CLI, régie theme and behaviour, which models and thinking levels agents may use, plugin enablement, safety limits, retention, and observability. Use for a guided question-by-question walkthrough of config.toml that discovers what it can from the machine and asks the rest in plain language, including first-time setup on a new machine.","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fskills\u002Fbuiltin\u002Ftheater-configure\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":18},"01M7AX36NADMJ5ZCAQC5T05ZM8","reviewer","Master of a pair play. Spawn one builder, review its work read-only, debate with evidence until you both agree, verify, and report. Attach this card when spawning a reviewer; never load it in your own chat.","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fskills\u002Fbuiltin\u002Fpair\u002Freviewer\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":18},"01M7AX34K85MAFMSA3CV0X8RW2","builder","Builder in a pair play. Implement the brief in the shared worktree, defend your choices with evidence, and iterate with the reviewer until you both agree.","https:\u002F\u002Fgithub.com\u002Fmana-byte\u002Ftheater\u002Fblob\u002Fmain\u002Ftheater\u002Fskills\u002Fbuiltin\u002Fpair\u002Fbuilder\u002FSKILL.md",[],[]]