@hybridlabor-api/aos 4.12.0 → 4.13.0

package/.agents/vendor-manifest.json

@@ -64,6 +64,12 @@
       "purpose": "Generator/validator pairs for GitHub Actions, Dockerfile, bash scripts and Makefiles, with reference libraries",
       "target_skill": "skills/global_config/{github-actions,dockerfile,bash-script,makefile}-{generator,validator}"
     },
+    "humanlayer-skills": {
+      "upstream": "https://github.com/humanlayer/skills",
+      "pinned_commit": "ca7c8088db69e315a8b2deea43820270457f8f3c",
+      "purpose": "Interview-driven generator for recurring agentic control loops in CI (manual, opt-in, not a pipeline node)",
+      "target_skill": "design-control-loop"
+    },
     "claude-code-templates": {
       "upstream": "https://github.com/davila7/claude-code-templates",
       "pinned_commit": "691832ad9c2506d15a8b752826135f0afa83d197",

package/.claude/agents/reviewer.md

@@ -1,7 +1,7 @@
 ---
 name: reviewer
 description: "Adversarial review of build-node output against the plan's stated contract, before Shipping runs its automated gate. Modeled on the `doubt-driven-development` discipline (see `.agents/graph.md`): reads the artifacts and the contract, never the implementer's claim that it's done, never their reasoning — passing the claim biases toward agreement. Prompted adversarially (\"find what is wrong,\" never \"does this look good\"). Classifies every finding by a fixed precedence (contract misread → valid & actionable → valid trade-off → noise). The dispatcher escalates instead of repeating a cycle if a repair round reports the exact same finding ID Reviewer already flagged (a no-progress guard — see `.agents/graph.md`'s Reviewer discipline for why this replaced the original \"two clean cycles\" framing)."
-model: o3-mini
+model: opus
 ---
 Adversarial review of build-node output against the plan's stated contract, before Shipping runs its automated gate. Modeled on the `doubt-driven-development` discipline (see `.agents/graph.md`): reads the artifacts and the contract, never the implementer's claim that it's done, never their reasoning — passing the claim biases toward agreement. Prompted adversarially ("find what is wrong," never "does this look good"). Classifies every finding by a fixed precedence (contract misread → valid & actionable → valid trade-off → noise). The dispatcher escalates instead of repeating a cycle if a repair round reports the exact same finding ID Reviewer already flagged (a no-progress guard — see `.agents/graph.md`'s Reviewer discipline for why this replaced the original "two clean cycles" framing).
 

package/.claude/agents/techlead.md

@@ -1,7 +1,7 @@
 ---
 name: techlead
 description: "Reviews Architect's plan for a capability map (module boundaries, dependency direction, build order) before any build node starts. Approves or rejects the plan back to Architect. Coordinates *what needs to happen*, not *who calls whom* — the dispatcher still does the actual invoking."
-model: gemini-3.8-flash-high
+model: sonnet
 ---
 Reviews Architect's plan for a capability map (module boundaries, dependency direction, build order) before any build node starts. Approves or rejects the plan back to Architect. Coordinates *what needs to happen*, not *who calls whom* — the dispatcher still does the actual invoking.
 

package/.opencode/agents/architect.md

@@ -1,7 +1,7 @@
 ---
 description: "Turns the user's goal (or `/bdbrainstorm` / `/grill-me` output) into a system plan. Reads existing architecture before proposing changes. Does not coordinate execution or invoke other agents — that is TechLead's job, decided by the dispatcher, not by Architect."
 mode: subagent
-model: opus
+model: opencode/muse-spark-1.3-contributor-free
 ---
 Turns the user's goal (or `/bdbrainstorm` / `/grill-me` output) into a system plan. Reads existing architecture before proposing changes. Does not coordinate execution or invoke other agents — that is TechLead's job, decided by the dispatcher, not by Architect.
 

package/.opencode/agents/reviewer.md

@@ -1,7 +1,7 @@
 ---
 description: "Adversarial review of build-node output against the plan's stated contract, before Shipping runs its automated gate. Modeled on the `doubt-driven-development` discipline (see `.agents/graph.md`): reads the artifacts and the contract, never the implementer's claim that it's done, never their reasoning — passing the claim biases toward agreement. Prompted adversarially (\"find what is wrong,\" never \"does this look good\"). Classifies every finding by a fixed precedence (contract misread → valid & actionable → valid trade-off → noise). The dispatcher escalates instead of repeating a cycle if a repair round reports the exact same finding ID Reviewer already flagged (a no-progress guard — see `.agents/graph.md`'s Reviewer discipline for why this replaced the original \"two clean cycles\" framing)."
 mode: subagent
-model: o3-mini
+model: opencode/muse-spark-1.3-contributor-free
 ---
 Adversarial review of build-node output against the plan's stated contract, before Shipping runs its automated gate. Modeled on the `doubt-driven-development` discipline (see `.agents/graph.md`): reads the artifacts and the contract, never the implementer's claim that it's done, never their reasoning — passing the claim biases toward agreement. Prompted adversarially ("find what is wrong," never "does this look good"). Classifies every finding by a fixed precedence (contract misread → valid & actionable → valid trade-off → noise). The dispatcher escalates instead of repeating a cycle if a repair round reports the exact same finding ID Reviewer already flagged (a no-progress guard — see `.agents/graph.md`'s Reviewer discipline for why this replaced the original "two clean cycles" framing).
 

package/.opencode/agents/techlead.md

@@ -1,7 +1,7 @@
 ---
 description: "Reviews Architect's plan for a capability map (module boundaries, dependency direction, build order) before any build node starts. Approves or rejects the plan back to Architect. Coordinates *what needs to happen*, not *who calls whom* — the dispatcher still does the actual invoking."
 mode: subagent
-model: gemini-3.8-flash-high
+model: opencode/muse-spark-1.3-contributor-free
 ---
 Reviews Architect's plan for a capability map (module boundaries, dependency direction, build order) before any build node starts. Approves or rejects the plan back to Architect. Coordinates *what needs to happen*, not *who calls whom* — the dispatcher still does the actual invoking.