agentleFS
loganware05loganware05/captains-compass-cursor/.cursor/skills/context-inodes/SKILL.md

context-inodes

Skill2 starsChanged 32 days agoMITView on GitHub

Builds and walks the content-addressed inode metadata store and derived context tree for filesystem-gated, module-scoped agent context

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this Skill when planning or reviewing work that touches module boundaries, when subagent context needs scoping to a module, or when the …

    line 8

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Context Inodes & Route Walking (M40)

Use this Skill when planning or reviewing work that touches module boundaries, when subagent context needs scoping to a module, or when the inode store is stale/missing.

Concepts

  • Inode — .agent/inodes/<sha256>.json: structural metadata for one source file (exported interfaces, function signatures with param optionality, I/O types, declared complexity), content-addressed and decoupled from raw source. Inodes never contain file contents.
  • Context tree — .agent/context/<domain>/<module>/node.json: derived from the source tree (never hand-authored). Each node carries a summary, child segments, and the inode refs for its subtree.
  • Route walk — sequential segment-by-segment resolution (src/lib → load src/node.json, then src/lib/node.json). Walks return inode pointers; callers page in source explicitly only when needed.

Declared complexity convention

Complexity is declared, never inferred. Annotate with @complexity:

def process(records: list[dict]) -> list[dict]:
    """@complexity O(N)"""
/** @complexity O(N) */
export function validateForm(fields: InputProps[]): string[] { ... }

Commands

# Build/refresh the inode store + context tree (deterministic)
./scripts/build-context-inodes.sh [--repo-root PATH]

# Fail (exit 1) when indexed sources are stale
./scripts/build-context-inodes.sh --check

# Walk a context route (prints steps + in-scope inode refs)
./scripts/walk-context-route.sh --route orchestrator/review

# Via the launcher
./scripts/northstar context build --repo /path/to/repo
./scripts/northstar context walk --repo /path/to/repo --route src/lib

Gates and invariants

  • doctor.sh fails closed when .agent/inodes/ exists but is stale (rebuild hint), and when committed Skill inodes are stale.
  • The northstar review boundary gate skips with an explicit note when the inode store is absent or stale — it never reviews against untrusted metadata. See Skill code-reviewer for the boundary checks.
  • Subagent manifests carry working_context (context root, inode refs, scope allow/deny) — see Skill capability-planning.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.