| compatibility | Python 3.8+, standard library only. No network. No third-party packages. |
| description | Produces a DSE pre-market daily briefing for a user's portfolio and watchlist using conditions-and-levels phrasing (never buy/sell commands). Use when the user asks for a morning briefing, pre-market summary, "what should I watch today", overnight news/disclosure recap, or a rundown of positions near stops/targets, watchlist names near entry, today's economic/earnings calendar, and risk items for Dhaka Stock Exchange holdings. |
| license | Apache-2.0 |
| metadata | {"author":"stock-buddy","github-path":"skills/daily-briefing","github-pinned":"main","github-ref":"refs/heads/main","github-repo":"https://github.com/kuntal-r-d/my-skills","github-tree-sha":"850204073caf1c909aebb10a9836e83c8d907f02","mode":["momentum","investment"],"prd_refs":["PRD-002:REQ-109","PRD-001:REQ-050"],"version":"0.1.0"} |
| name | daily-briefing |
Daily Briefing
When to use
Activate when a user wants a pre-market read on their book: "give me my morning briefing",
"what's near my stops today?", "any overnight news on my holdings?", "watchlist names close
to entry?", "what's on the DSE calendar today?". This is the Daily Briefing Agent
(PRD-002 REQ-109; PRD-001 REQ-050). It composes the outputs of other skills' levels into a
single readable card on a post-close / pre-market dual schedule (see references).
Critical tone rule
The briefing never issues commands. It describes conditions and levels only:
"GP is within 2% of its stop level (price 178.0 vs stop 182.0)" — not "sell GP".
Every generated line passes a no-imperative guardrail (strip_imperatives in
scripts/brief.py) that rewrites or flags command verbs ("buy", "sell", "watch for",
"if X then consider"). See references/BRIEFING.md.
What it does
Builds six sections from the input:
- Market regime — one line from
macro_regime (rating + risk multiplier).
- Held positions near levels — positions whose
current_price is within ~3% of a
stop_level or target_level.
- Watchlist near entry — watchlist names within ~3% of their
entry_level.
- Calendar today — economic/earnings events dated
as_of.
- Overnight news / disclosures — headlines mapped to held/watch tickers.
- Risk items — circuit/floor/halt mentions and single-name concentration.
How to run
python3 scripts/brief.py --input briefing_input.json --pretty
cat briefing_input.json | python3 scripts/brief.py
Input fields: as_of, optional user, portfolio.positions[]
(ticker, qty, avg_cost, current_price, stop_level?, target_level?), watchlist[]
(ticker, current_price, entry_level?, signal?), calendar[] (date, event),
overnight_news[] (ticker, headline, source), macro_regime (rating, risk_multiplier).
Interpreting output
Returns summary (1–2 sentences), markdown (the full briefing), sections{...},
item_counts{}, and flags[]: stale_briefing (missing/old as_of), fallback
(no positions supplied). All phrasing is conditional — treat as situational awareness,
not instructions. Educational analysis only, never financial advice.
