Menu

Individual skill

asteria-fleet-dq-audit

0upvotes

Reconcile and certify a data collection in the Asteria Fleet Data Quality Hub. Deduplicate overlapping source snapshots (CERTIFIED wins), normalize measures with reference aliases/conversions/fx, classify mismatches vs. unrecognized/ambiguous vs. quarantined, compute normalized totals and rankings, infer internal control codes per scoped public ID, and emit ONE answer JSON conforming exactly to the task's answer_template.json. Use whenever a task points at <TASK_ENV_BASE_URL> / the Asteria Fleet Data Quality Hub and asks for a reconciled audit plus a certification/close decision returned as a single JSON object.

experiments/claude_code_glm_5_2_max/reports/skills/task_group_021/fewshot/fewshot_attempt_02/SKILL.md

Read pinned source

Source commit 5082a777379156a048881e94940d640702e5e09a

What was checked

Matches the observed repository HEAD
Yes
Scanner rules are current
No
Repository scan complete
No
Skill scan complete
No
Observed behavior
Not checked yet
Last checked
Nov 6, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings)
121
Repository last updated
Oct 29, 2026
License
Apache-2.0

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files
Read excerpt

# Asteria Fleet Data Quality Hub — Reconciliation & Certification Audit

This skill solves the recurring task family: *"audit collection X as of cutoff Y against the Asteria Fleet Data Quality Hub and return one certified JSON answer."* Every task in the family has the same shape — only the collection, cutoff, focus IDs, thresholds, and output contract change. The procedure below is parameterized by those inputs; it does **not** depend on any specific answer. Never copy values from another task's answer — re-derive everything for the current task's data.

## Inputs (always present in the task directory)

- `prompt.txt` — narrative of what to audit and report.
- `payloads/case_scope.json` — the **parameters**: `collection_id`, cutoff timestamp, focus/decision IDs, ranking limits and ordering rules, certification thresholds, status→action map, control-case anchors.
- `payloads/answer_template.json` — the **output contract**: a JSON Schema (or field contract) with required keys, `enum`s, `pattern`s, `minItems`/`maxItems`, ordering rules, and numeric precision. Emit EXACTLY this shape.
- `environment_access.md` — the **only** source of network access: base URL, bearer token, and the allow-list of endpoints.

## Golden rules (read first)

1. **One JSON object, exactly the template.** No extra keys, no missing keys, no commentary, no Markdown fences. Templates use `additionalProperties: false` — every key you emit must be in the schema.
Back to top