spl-to-apl

Warn

Audited by Runlayer on Feb 22, 2026

Risk Level: MEDIUM
Scan Summary
Max Score
78%
Files
11
Flagged
9
Chunks
15
Flagged Files (9)
.meta/2026-01-28-schema-read-findings.mdHIGH
78.3%

Malicious tool definition detected

Tool: .meta/2026-01-28-schema-read-findings.md Description: --- date: 2026-01-28 thread: https://ampcode.com/threads/T-019c0497-6c51-7113-a9a0-bf9ea2efe9e2 branch: feat/spl-to-apl-v2-skill status: paused keywords: - skill-design - schema-read - eval - progressive-disclosure --- # schema-read behavior findings attempted to improve model's schema-read rate (calling readFile on reference/dataset-schemas.md before translating).

.meta/DESIGN-NOTES.mdHIGH
78.3%

Malicious tool definition detected

Some Axiom datasets store numeric-looking fields as strings (e.g., `status` in sample-http-logs). This causes silent failures: - `status >= 500` compiles and runs - But does string comparison, returning wrong results - Correct: `toint(status) >= 500` The type information is dataset-specific and lives in `reference/dataset-schemas.md`. We want the model to read this file before translating.

.meta/cases.tsHIGH
78.3%

Malicious tool definition detected

Tool: .meta/cases.ts Description: import type { TranslationCase } from "../../../eval-tooling/src/shared/types"; /** * Test cases validated against Axiom Playground (play.axiom.co) with real datasets.

.meta/spl-to-apl.eval.tsHIGH
78.3%

Malicious tool definition detected

Tool: .meta/spl-to-apl.eval.ts Description: import { config } from "dotenv"; import { fileURLToPath } from "node:url"; import { dirname, resolve } from "node:path"; const __filename = fileURLToPath(import.meta.url); const __dirname = dirname(__filename); // Load .env from eval-tooling directory config({ path: resolve(__dirname, "../../../eval-tooling/.env") }); import { Eval, Scorer } from "axiom/ai/evals"; import { testCases } from "./cases"; import { flag, pickFlags, getGitCommit, buildSkillMe

README.mdHIGH
78.3%

Malicious tool definition detected

Tool: README.md Description: # spl-to-apl Translates Splunk SPL queries to Axiom APL.

SKILL.mdHIGH
78.3%

Tool passed security scan

Malicious tool definition detected

Description: `reference/function-mapping.md` — complete function list - `reference/examples.md` — full query translation examples - APL docs: https://axiom.co/docs/apl/introduction

tests/test-queries.mdHIGH
78.3%

Malicious tool definition detected

Tool: tests/test-queries.md Description: # SPL to APL Test Queries Test the skill by translating these SPL queries to APL and running in [Axiom Playground](https://play.axiom.co). ## Test Cases for sample-http-logs ### 1.

reference/examples.mdMEDIUM
68.9%

Tool passed security scan

reference/command-mapping.mdMEDIUM
68.2%

Tool passed security scan

Passed Files (2)Click to expand
reference/dataset-schemas.mdOK
5.5%

Tool passed security scan

reference/function-mapping.mdOK
5.5%

Tool passed security scan

Audit Metadata
Max File Score
78%
Classification
UNKNOWN_SERVER
Files Scanned
11
Files Flagged
9
Chunks Analyzed
15
Analyzed
Feb 22, 2026, 04:47 AM