Menu

Individual skill

pho-observatory-audits

0upvotes

Solve Public Health Observatory web-portal audit tasks requiring exact JSON outputs from analysis_request.json and answer_template.json. Use for PHO algorithmic publication audits over state, county, or country data, including exact protocols PHO_STATE_TRANSPORT_AUDIT_V1, PHO_COUNTY_MEDIATION_TRANSPORT_V1, PHO_STATE_ROBUSTNESS_TRANSPORT_V1, PHO_COUNTY_PANEL_TRANSPORT_V1, and country burden revision/PCA/cluster/panel audit requests.

experiments/codex_skill_creator_comparison_gpt5_5_xhigh/reports/skills/task_group_023/fewshot/codex/fewshot_attempt_02/SKILL.md

Read pinned source

Source commit 5082a777379156a048881e94940d640702e5e09a

What was checked

Matches the observed repository HEAD
Yes
Scanner rules are current
Yes
Repository scan complete
No
Skill scan complete
No
Observed behavior
Network or broad shell
Last checked
Nov 11, 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

# PHO Observatory Audits

## Core Workflow

1. Read the task prompt, `analysis_request.json`, and `answer_template.json` before fetching evidence.
2. Resolve `<TASK_ENV_BASE_URL>` from the task environment and use only the public portal endpoints. Read [portal_schema.md](references/portal_schema.md) before downloading data.
3. If `analysis_request.json` has one of the exact protocol IDs listed below, read the matching section in [protocol_profiles.md](references/protocol_profiles.md). Do not activate a profile by similar subject matter or family name.
4. Freeze one effective request contract before computation. Apply direct keys and `_overrides` aliases exactly as the active profile says; arrays replace whole arrays, objects merge recursively, and unknown targets fail before evidence resolution.
5. Fetch only the required datasets. `scripts/fetch_portal.py` can download CSV files from the portal into a local working directory.
6. Resolve releases independently for every requested source, measure, geography, and year. Select records using the request's ordered revision/release/id priority; suppressed, invalid, withdrawn, blank, or null analytic values are unavailable and are never zero-filled.
7. Build every cohort from the effective completeness predicates. Preserve all declared entity, time, feature, fold, grid, checkpoint, cluster, and output orders. Never independently sort aligned result arrays.
8. Complete every module before evaluating controlled decisions. Evaluate gates on unrounded values, then round only at final JSON serialization.
9. Return exactly one JSON object conforming to the template. Emit no narrative, comments, NaN, Infinity, or extra keys unless the template explicitly requires them.
Back to top