context-combiner

Merges multiple confirmed requirement analyses into one deduplicated, conflict-resolved specification.

6|1|Updated Jul 4, 2026
One-click install
npx skills add https://github.com/azure-id/orc --skill context-combiner-azure-id
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: context-combiner
Source: https://github.com/azure-id/orc/tree/main/templates/skills/context-combiner
Command: npx skills add https://github.com/azure-id/orc --skill context-combiner-azure-id

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When you analyze several related requirement documents separately, you end up with overlapping, duplicated, or contradictory requirement specs that no planner can build from directly. This Skill merges 2+ already-confirmed ORC analyses into a single reconciled requirement context, proving that no source requirement was lost in the merge. ## Core Features & Use Cases - Requirement reconciliation: Pools all source requirements into one table and classifies them as exact duplicates, semantic duplicates, partial overlaps (split, never collapsed), conflicts, or ordering dependencies. - Interactive conflict resolution: Raises every cross-source conflict as a one-at-a-time challenge with a recommended option, checkpointing each decision to combine-decisions.md so nothing is re-asked after a resume. - Conservation gate: Builds a source coverage matrix where every source requirement ID gets exactly one outcome, and blocks handoff until coverage reaches 100%. - Use Case: You analyzed three related feature documents for the same reporting module and chose "pass to context-combiner" at orc-analyze Phase F. The combiner verifies the analyses genuinely overlap, merges them into combined-report.md and combined-requirement-spec.md, and hands the merged spec to the planner unchanged. ## Quick Start Ask the orchestrator to combine the confirmed analyses for the reporting module into a single merged requirement spec before building.

Frequently Asked Questions about context-combiner

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I combine multiple requirement analyses into one spec?▼

Dispatch the context-combiner with the paths of 2 or more confirmed requirement-spec files. It pools all requirements, deduplicates and reconciles them, then writes a combined-report.md and a combined-requirement-spec.md that the planner consumes unchanged.

What happens if the analyses I want to merge are not actually related?▼

The combiner verifies real overlap through shared files, overlapping requirements, and shared scope. If overlap is weak or empty, it challenges you to combine anyway or keep them separate, and stops without writing any combined artifact if you choose separate.

How does the combiner prove no requirements were lost during the merge?▼

It builds a source coverage matrix where every source requirement ID gets exactly one outcome: merged, deduped-into, split-across, conflict-resolved, or dropped with a recorded user decision. Handoff is blocked until coverage reaches 100 percent.

Can the context-combiner build the merged requirements directly?▼

No. The combiner never builds and never spawns subagents. It returns the combined spec to the orchestrator, which offers to stop with the report or pass the spec to the build pipeline starting at the requirement planner phase.

What happens when merged evidence anchors are stale?▼

The combiner spot-checks each inherited file:line anchor against the current repository state. Failed checks mark the evidence STALE, and stale evidence on a buildable requirement triggers a challenge to re-anchor, proceed anyway, or drop it.

When should I not use the context-combiner?▼

Do not use it for unrelated analyses, single analyses, or mini lanes, since orc-analyze-mini and orc-mini do not support it. With more than 4 sources, merge quality degrades and staged combining is recommended instead.