Install
$ agentstack add skill-horizonbrute-standardized-ai-looping-language-saill-dirtree-check ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Skill: /dirtree-check
Model preference: #midcost
Verify that every instrumented directory in tested_implementations/ has a nonce and correct agents.md inheritance chain. Output a rendered tree to the documentation root and stamp it with a SHA256 hash.
Step 1 — Discover directories
find working_copy/tested_implementations -name "testing_nonce.md" -type f
Or use Glob: **/testing_nonce.md under working_copy/tested_implementations/.
Collect the parent directory of each hit. That is the full target list.
Step 2 — Collect nonce + agents.md chain per directory (parallel)
For each directory, spawn a haiku sub-agent with this brief:
> Read /testing_nonce.md — extract the value after nonce: . > Read each agents.md file in the inheritance chain (the directory itself, then each > parent directory up to the impl root that has an agents.md). From each, extract the > first line after # This context loaded from: . > Return ONE CSV row, no header, no other output: > directory_shortname,trace1|trace2|...,nonce > Traces ordered deepest-first.
Pre-compute the inheritance chain for each directory and pass it explicitly — do not ask the haiku agent to discover it. Walk the path from the target directory up to the impl root, collecting any directory that has an agents.md.
Spawn all sub-agents in a single parallel batch (one Agent call per directory, all in one message). Collect all CSV rows before proceeding.
Step 3 — Build the directory tree markdown
From the collected rows, render a markdown fenced code block using tree characters (├──, └──, │). For each node show:
- directory name, right-aligned nonce label
nonce: ✓ agents.md chain ( level(s)):with each trace on its own indented line
Group by implementation: impl 1, impl 2 (full hierarchy), impl 3.
Append a summary table:
| Impl | Directories checked | Max inheritance depth | All nonces confirmed |
Step 4 — Write output file
Write the complete tree to:
working_copy/documentation/lastrun_dirtree_check.md
Include at the top:
# Directory Tree Context Check
Run: | haiku agents | working_copy/tested_implementations
Step 5 — SHA256 hash
After writing the file, compute its SHA256:
python -c "import hashlib; d=open('working_copy/documentation/lastrun_dirtree_check.md','rb').read(); print(hashlib.sha256(d).hexdigest())"
Append to the bottom of the file:
---
sha256:
Then recompute and verify the hash matches before reporting done. If the file was modified by the append, the final hash is the hash of the file with the sha256 line included — write a placeholder, compute, replace, verify once.
Step 6 — Report
Print a one-line confirmation:
dirtree-check complete — /15 nodes ✓ sha256: ...
Do not print the full tree to the conversation — it is in the output file.
Notes
- The inheritance chain for impl 2 stacks up to 4 levels deep at project leaves. Impl 3
directories are siblings with no parent chain. Impl 1 is a single-level flat directory.
- If any nonce file is missing or any agents.md trace line is absent, flag the directory
as MISSING in the tree and note it in the report.
- The SHA256 is of the final file bytes including the
sha256:line itself. This is a
self-referential hash — compute it by writing the placeholder, computing, and replacing in one edit so the file is never re-read after hash insertion. Practical approach: write the file without the hash line, compute hash, append the line. The hash then covers the full file content.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: HorizonBrute
- Source: HorizonBrute/StandardizedAILooping_Language-SAILL
- License: MIT
- Homepage: https://horizonbrute.github.io/StandardizedAILooping_Language-SAILL/
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.