Read the applicable Context Tree before acting. In BYO sessions, route only among locally authorized Teams by reading each exact root SCOPE.md before selecting one task snapshot; in managed workspaces, use the bound Tree. Do not use for a Context Tree PR/MR review or an explicit broad audit of stored tree content.
Scanned 9/2/2026
Install to Claude Code
npx -y skills add first-tree-ai/first-tree --skill first-tree-read --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of First Tree Read?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/first-tree-ai-first-tree-read)More formats (shields.io, HTML) on the badges page.
---
name: first-tree-read
version: 0.8.1-local.1
description: Read the applicable Context Tree before acting. In BYO sessions, route only among locally authorized Teams by reading each exact root SCOPE.md before selecting one task snapshot; in managed workspaces, use the bound Tree. Do not use for a Context Tree PR/MR review or an explicit broad audit of stored tree content.
---
# First Tree Read — Local Context
## Purpose and authority
Read the current Agent's private Local Context before acting on a task with a
repo, path, feature, domain, owner, bug, or error signal. This payload is the
managed Local variant of `first-tree-read`; its public name and routing
description intentionally match the Remote payload.
Local Context is host-local truth for this Agent only. It is not Team-approved
truth, a Git checkout, a BYO snapshot, or a fallback for an unreadable or
invalid remote binding. `first-tree-seed`, `context-tree-review`, and
`context-tree-audit` never operate on it.
Apply the generated Context Tree Policy's normal/archive/member authority and
code-vs-tree drift rules. Never modify Tree content with this Skill.
## Workflow
1. Run the generated briefing's exact First Tree CLI invocation:
```text
<firstTreeInvocation> --json tree local resolve --ensure --intent read
```
Use only the returned `data.path`. The guard accepts no caller-supplied live
path and rechecks runtime identity, fixed Workspace containment, resource
limits, the remote-observed latch, and the live Server binding. Missing,
corrupt, unknown, invalid, frozen, or bound state stops the Local read.
2. Before reading business Context, run:
```text
<firstTreeInvocation> tree verify --tree-path "<live-root>"
```
A non-zero result stops the read. Report that Local Write must repair the
live Tree; do not guess an older version or continue from partial files.
3. Inspect the filesystem hierarchy command from the live root, then use its
selectors before native Markdown reads:
```text
<firstTreeInvocation> tree tree --tree-path "<live-root>" --help
<firstTreeInvocation> tree tree --tree-path "<live-root>" [path] [selectors]
```
This mode performs no Git discovery, pull, branch, commit, snapshot, forge,
Cloud Tree IO, or attribution. Keep the read focused: root and relevant
parents, matched leaves, and material `soft_links` only.
4. If a selected file vanishes, fails to parse, or appears to be an in-flight
concurrent edit, discard that read and restart from verification. Local
reads have no stable snapshot guarantee.
5. Immediately before using Local Context for a substantive choice, run the
same full `tree verify --tree-path` again. Passing proves only that the live
Tree is mechanically valid at that moment; it does not prove the bytes read
earlier were unchanged.
6. Carry forward only relevant durable decisions, constraints, ownership, and
cross-domain relationships. Local V0 has no immutable commit attribution,
so do not emit the standard remote Context Tree influence note or fabricate
source links.
## Failure boundary
Local writers use no content lock, snapshot, candidate, fingerprint, journal,
rollback, or per-write approval. A reader may observe moving state. Never turn
that accepted limitation into a claim of transaction safety or silently accept
an invalid intermediate Tree.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!