Install
$ agentstack add skill-yassimba-loom-quick-lineage ✓ 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.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Quick Lineage
One colored call graph in chat, grounded in stackdiff (AST-verified for TS/TSX, Python, Go, Rust) — the fast sibling of lineage and blueprint, for when the picture matters more than the ceremony.
1. Read the branch from the conversation
Infer both the entry (the function, command, or object's method the conversation centers on — with several candidates, the one whose call tree covers the most of what was discussed) and the branch:
- Discussed — the user is asking about something as it is → one world, current source.
- Landed — changes exist (session edits, working tree, a named ref) → diff the worlds.
- Planned — the change is still intent → real BEFORE rails plus a projected
±overlay.
State the choice in one line ("quick lineage of run, landed changes vs HEAD") before running anything. Done when: the entry and branch are named in chat — silently picking is how the wrong graph gets drawn.
2. Run stackdiff
Start at --max-depth 2 and deepen only the limb the question lives in — full depth buries the point.
| Branch | Command | | --- | --- | | Discussed | stackdiff --tree -e --max-depth 2 --format markdown (add ` for a past world) | | Landed | stackdiff -e --max-depth 2 --only-changes --format markdown — working tree included; add for ref-to-ref | | Planned | the Discussed command for BEFORE rails, then hand-write +/-` lines for the intended calls |
--format markdown emits the ``` `diff `` fence ready to paste; rich mode annotates each node with binding = call(args) → return, its doc line, and path:line. Unchanged limbs trim to … by default (widen with --context N, --full for everything). When the question is convergence — "do these paths reach the same thing?" — use --view lineage`: the call DAG at data granularity, each function drawn once with fan-in visible, data objects as stadium nodes, bindings riding the edges.
Entry not found → stackdiff --tree lists the exported entrypoints; pick the nearest and say you did. Language outside TS/TSX/Python/Go/Rust → say stackdiff cannot parse it and trace by hand, marked as unverified. "Who calls X" questions → --callers -e X inverts the graph.
Tuning noise
Builtin/plumbing calls (len(), clone(), str(), println!) hide by default so the graph is your code talking to your code. When a graph still reads cluttered, tighten it in one loop: run stackdiff --noise-report (a frequency table of unresolved calls, marked hidden/shown), add hide = ["glob", …] globs for the framework plumbing you see to .stackdiff.toml at the repo root (show = […] rescues false positives), and re-run. --noise reveals everything when you suspect the filter ate signal.
3. Show it colored
Paste the markdown output as-is — the ``` `diff `` fence renders + green and - red (!` marks a same-call-different-arguments change), and the two-space status column keeps rails aligned. For Planned, projected lines are the design's promise, not source facts — say so above the block.
Under the block, one takeaway sentence: what the graph shows about the question that prompted it.
Done when: the graph is in chat with every +/- line either machine-verified by stackdiff or explicitly marked projected, and the takeaway names what changed or will change.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Yassimba
- Source: Yassimba/loom
- License: Apache-2.0
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.