Asteria Portfolio JSON
Workflow
- Read the request payload and answer template first.
- Treat current Asteria environment records as the source of truth. Use local payloads and memos only as task framing when they conflict with live records.
- Read only the needed endpoint families. See workflow notes for the endpoint map and calculation rules.
- Match the template exactly. Keep required keys, enum values, ordering, and numeric precision unchanged.
- Use the current environment
as_of_dateand policy identifiers, not request dates, memo dates, or stale worksheet snapshots. - Return only the JSON object.
Output Checks
- Preserve the requested order for rows, pairs, trades, and sleeves.
- Round numeric fields to the precision declared by the template.
- Keep ids alphabetized when the template asks for alphabetical ordering.
- Do not add commentary, markdown, or extra keys.
- Do not copy stale sample values into the answer.
Task Families
- Credit trade packages and rotations: use current holdings, bond metadata, issuer watchlist status, and portfolio constraints to select eligible names and size trades.
- Correlation reviews: compute Pearson correlations from monthly simple returns over the requested window, then identify the strongest positive pair and the weakest pair in the stated universe.
- Allocation views: combine current macro signal scores with prior views to set
view,change,conviction, andrationale_code, plus any required overlay fields. - Committee summaries: combine the correlation and allocation outputs, then choose the trigger and next step that match the concentration story.