Run three parallel lenses to propose file-specific solutions or escalate to L2.
npx skills add https://github.com/Kohlex/autopilot-plugin --skill escalation-l1
Please help me install this Agent Skill. Skill: escalation-l1 Source: https://github.com/Kohlex/autopilot-plugin/tree/main/skills/escalation-l1 Command: npx skills add https://github.com/Kohlex/autopilot-plugin --skill escalation-l1
The L1 escalation skill autonomously responds to in-architecture obstacles by spawning three parallel subagents (Reuse, Compose, Extend-one-function) to explore alternative execution strategies without altering the locked architecture.
Trigger the L1 escalation using the /escalation-l1 command inside the locked architecture.
High-intent search queries and answers about installing and using this skill.
To resolve in-architecture obstacles, run three parallel subagent lenses that explore reuse, compose, and extend-one-function strategies without altering the locked architecture, outputting concrete proposals.
Autonomous L1 escalation is a parallel exploration process needed when straightforward architectural changes fail, spawning three subagents to evaluate risk, effort, and acceptance criteria for alternative execution strategies.
Trigger autonomous escalation within a locked architecture by using the /escalation-l1 command, which initiates the parallel subagent lenses to generate one to three proposals tied to specific files.
Yes, you can use parallel subagent lenses for obstacles within plan and spec constraints, as the skill targets issues that cannot be solved with straightforward in-architecture changes while respecting existing limits.
When all parallel architecture lenses fail, the synthesis step escalates the obstacle to L2 and updates the capability plan, ensuring unresolved in-architecture obstacles receive higher-level intervention.
The best way to synthesize alternative execution strategies is a dedicated synthesis step that evaluates risk, effort, and acceptance-criteria coverage across all subagent proposals to select the best-fit path.
HechaoYannetGenerate a master architecture document from GDDs, ADRs, and governance constraints.
SpaceZephyrAnalyze weekly memory records and extract recurring behavior patterns for L2 promotion.
marcusgollGenerate sprint-plan.md with dependency graphs and locked OpenAPI contracts.
VedaAstroOrchestrate autonomous subagent delegation for parallel workstreams and independent review.
