Individual skill
fleet-data-quality-reconciliation
Reconcile Asteria-style fleet data-quality hub collections into schema-valid audit JSON. Use when a task provides a prompt, case scope, answer contract/template, and read-only hub access for contact rosters, fuel or freight ledgers, maintenance histories, duplicate resolution, source-snapshot reconciliation, alias classification, unit/FX normalization, readiness counts, quarantine sets, ranked exception panels, or compact internal control-code panels.
experiments/codex_gpt5_5_xhigh/reports/skills/task_group_021/reflect-3/reflect-3_attempt_02/SKILL.md
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.
- agent Repository topic
- ai Repository topic
- benchmark Repository topic
- llm Repository topic
- self-evolving Repository topic
- self-learning Repository topic
- skills Repository topic
- legal Derived category
- data Derived category
- security Derived category
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
# Fleet Data Quality Reconciliation
## Core Workflow
1. Read the prompt, case scope, and answer contract before querying data. Extract the collection ID, cutoff, business period, scoped stable IDs, required sort orders, numeric precision, allowed enum values, and whether totals include or exclude exceptions.
2. Inspect catalog/schema and source snapshots for the scoped collection. Treat snapshot business cutoffs and statuses as source-selection evidence; do not filter source rows by ingestion time unless the task explicitly says to.
3. Pull all in-scope rows through the read-only data interface. Use structured queries over paginated endpoint sampling when the collection may exceed one page.
4. Build a reproducible reconciliation table with raw rows, retained logical records, normalized fields, exception flags, and reason flags. Derive every output field from that table.
5. Validate the final object against the answer contract: exact top-level keys, no extra commentary, sorted arrays, unique IDs, integer counts, required enums, and stated rounding.
## Source Resolution
- Prefer certified snapshots over provisional snapshots when the same stable public ID appears more than once. Keep provisional-only logical records unless the prompt says to certify only an authoritative snapshot.
- Count `raw_row_count` from in-scope raw rows, `logical_*_count` from retained stable IDs, and duplicate raw count as raw rows minus retained logical records when duplicates are defined by the stable ID.