Skip to content

BuilderIO/agent-native / .agents/skills/context-xray

context-xray

Verified

Inspect and manage the live agent context window with Context X-Ray. Use when context is getting large, the user asks what is in context, or stale tool results/files should be pinned, evicted, restored, or reported by an external host.

Context X-Ray

Context X-Ray is the framework's context garbage-collection surface. It shows the current thread's model-bound context as content-derived segments with token counts, then lets the user or agent pin, evict, or restore individual segments.

Actions

ActionWhen to use
context-manifest-getRead the current manifest for a thread. Returns token totals, segment status, source, and whether changes are enforceable.
context-preview-getCompose the would-be system context for the current user/org/app without a live thread. Returns labeled system sections with provenance, governance tier, token counts, and bounded previews. Backs the Agent page Context tab.
context-pinPreserve a segment across future compaction/model calls. Use for task specs, acceptance criteria, user constraints, and other durable context.
context-evictExclude a stale or irrelevant segment from future model calls. Eviction is reversible and never deletes chat history.
context-restoreUndo a pin, evict, or summarize directive for a segment.
context-reportExternal hosts can report their visible context inventory. These manifests are advisory unless Agent-Native owns the emitted content.
context-compactCompact the thread now: fold every not-yet-compacted message into observational memory instead of waiting for the post-turn threshold. Takes effect on the next model call; the transcript is unchanged.

System Sections

Manifests now cover the system-prompt half of context, not just conversation segments. Each system section carries a provenance label (framework-core, actions-prompt, template, enterprise-workspace-core, sql-workspace, legacy-app-default, organization, personal, memory, db-schema, tools, model-overlay, runtime-context) and a governance tier:

  • required — framework/enterprise policy; can never be altered by the user.
  • inherited — template/org/workspace instructions; managed elsewhere, visible here.
  • user — personal instructions and memory; editable through Files.

System sections are display-only provenance: they cannot be pinned, evicted, or summarized. The persisted manifest stores source refs, scope, content hashes, token counts, and bounded previews (~200 chars) — never raw prompt bodies; full content stays behind the resource/access model. The in-chat X-Ray panel groups these under "System", and the Agent page Context tab renders the full breakdown (via context-preview-get) with a system-vs-conversation split meter. Manifests are the latest snapshot per thread; there is no per-iteration history.

Rules

  • Never evict or summarize protected segments. The manifest marks active-turn user/tool/thinking context as protected.
  • System sections are never evictable; required/inherited tiers are locked by design — do not present controls that imply otherwise.
  • Prefer pinning the user's task, requirements, and decisions before evicting large stale tool results.
  • Eviction excludes content from future model calls; it does not delete the canonical transcript or files.
  • In external/advisory mode, be honest: recorded directives are intent for the host except for Agent-Native-originated content we can actually withhold.
  • If token counts are estimated, describe reclaim as approximate.

Typical Flow

  1. Call context-manifest-get with the active threadId.
  2. Sort segments by tokenCount and inspect large stale Tool results or Files read entries.
  3. Call context-pin for essential specs or user instructions.
  4. Call context-evict for large irrelevant segments.
  5. Offer context-restore if the user wants undo.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 319kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 319kUpdated Claude CodeCodexVerifiedInstall