Install
$ agentstack add skill-arasz-ai-badger-debug-issue ✓ 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
Debug issue
Trace the call chain before hypothesizing. The failure this skill exists to prevent: proposing a fix before the call chain from symptom to entry point has actually been traced.
The workflow derives from the skill templates the code-review-graph project auto-installs (MIT, © 2026 Tirth Kanani), rewritten here to be tool-agnostic: every step carries a baseline that needs no graph server.
Steps
- Locate the code from the symptom.
Accelerated: a code-graph semantic-search call (e.g. semantic_search_nodes-shaped; discover the real tool name from the server's own listing) finds relevant code by meaning, not string match. Baseline: grep for the error message, log line, or another distinctive identifier (field name, status code, exception type) across the source tree.
- Trace callers and callees from that point — both directions, not just one.
Accelerated: a graph query for callers-of/callees-of the located symbol (e.g. query_graph- shaped with callers_of/callees_of patterns). Baseline: the IDE's find-references / go-to-definition, or grep -rn the symbol for callers and read the body for callees. Walk outward until reaching a test that exercises this path or a public entry point (HTTP handler, CLI command, message consumer, orchestration trigger).
- Follow the full execution path through the suspected area — the entry point that triggers
the bug is often several hops from where the symptom surfaces. Accelerated: a flow query (e.g. get_flow-shaped) returns the whole path in one call. Baseline: read each hop in sequence, noting where state changes or a branch could diverge.
- Check whether a recent change introduced it.
Accelerated: a change-detection call (e.g. detect_changes-shaped) flags recent risk-scored changes near the traced path. Baseline: git log -S'' --oneline -- or git blame the suspected lines — recent changes are the most common source of new issues, check this before assuming the bug is old.
- Only now form a hypothesis, stated as "execution reaches line X via path Y, and the bug is
Z because of change W" — or "the bug predates commit V because no recent change touched this path". A hypothesis with no traced path behind it is a guess, not a diagnosis.
Gotchas
No environment-specific gotchas known.
Red flags — STOP
- Proposing a fix before step 2 has produced an actual caller/callee list
- Tracing only one direction — callers alone miss what the function itself calls into
- Treating "similar code exists elsewhere" as tracing this call chain
- Skipping step 4 — a large share of new bugs are yesterday's diff; the history check is cheap
relative to guessing
A diagnosis is not finished until it names the specific call path (caller → … → entry point) showing the bug is reachable, and states whether a specific recent change is implicated. "I think it's this function" with no traced path is not a diagnosis.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Arasz
- Source: Arasz/ai-badger
- License: MIT
- Homepage: https://github.com/Arasz/ai-badger
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.