← Samaya Finance SkillsCONTENT HISTORY

Update to Samaya Finance Skills

Snapshot Nov 16, 2026 · 18:17 UTC · version 1.2.2

Collection source: downloaded plugin package.

WHAT CHANGED

First saved snapshot

No earlier snapshot is available to establish a change.

Save Samaya Finance Skills and check future changes in Following.

Up to 3 plugins free · no card required.

Compare saved observations

Download comparison JSON
Full technical diff · 0 changed fields
Full snapshot data
{
  "description": "\"annotated price chart\", \"what moved the stock\", \"price performance with catalysts\", \"label the chart with the news\" — for meeting, pitch or position-review prep: an interactive price-volume chart for one stock, split into periods, each big move and run labelled with the news that drove it, sourced via Samaya. Not for price alerts (samaya-trigger-price-move), co-moves (samaya-correlation-analysis) or insider trades.",
  "included_files": [
    {
      "relative_path": "assets/template.html",
      "size_in_bytes": 51101
    },
    {
      "relative_path": "references/attribution.md",
      "size_in_bytes": 9353
    },
    {
      "relative_path": "references/data-pulls.md",
      "size_in_bytes": 6474
    },
    {
      "relative_path": "references/events-schema.md",
      "size_in_bytes": 6861
    },
    {
      "relative_path": "references/example-events-aapl.json",
      "size_in_bytes": 20544
    },
    {
      "relative_path": "scripts/compute.py",
      "size_in_bytes": 17005
    },
    {
      "relative_path": "scripts/coverage_check.py",
      "size_in_bytes": 3048
    },
    {
      "relative_path": "scripts/integrity.py",
      "size_in_bytes": 6510
    },
    {
      "relative_path": "scripts/phases.py",
      "size_in_bytes": 2299
    },
    {
      "relative_path": "scripts/plan.py",
      "size_in_bytes": 1815
    },
    {
      "relative_path": "scripts/render.py",
      "size_in_bytes": 24979
    }
  ],
  "name": "samaya-annotated-price-performance",
  "skill_md_contents": "---\nname: samaya-annotated-price-performance\ndescription: >-\n  \"annotated price chart\", \"what moved the stock\", \"price performance with catalysts\", \"label the\n  chart with the news\" — for meeting, pitch or position-review prep: an interactive price-volume\n  chart for one stock, split into periods, each big move and run labelled with the news that\n  drove it, sourced via Samaya. Not for price alerts (samaya-trigger-price-move), co-moves\n  (samaya-correlation-analysis) or insider trades.\n---\n\n# Annotated price performance\n\n**What it includes:** an interactive price chart split into periods, each big move labelled with\nthe news or event that drove it, with sources.\n\nProduces one self-contained interactive HTML page for one stock over a window (default: the\ntrailing 12 months to the last close):\n\n- a one-sentence takeaway and 2 to 5 phase tiles, each titled with its own takeaway;\n- a price chart (default view) with the benchmark rebased, numbered markers on every day the\n  stock moved well beyond what the market explains, and open triangles for major company events\n  that did not move it;\n- a volume pane, because volume tells as much of the story as price;\n- a daily split view (each day's move into a market part and a stock-specific part, against the\n  stock's normal range) and a cumulative split view;\n- a side panel per event (what happened, attribution tier, the split, sources), an event table,\n  and collapsed sections for market-driven days, method and sources.\n\nThe design principle: **the numbers decide which days get a story; research decides what the\nstory is; a validator decides whether the page may be written.** Earlier versions of this\noutput invented causes for drift, pinned macro headlines on market days and missed timing. Every\nrule below exists because of one of those failures, and most are enforced in code.\n\n## Before starting\n\nConfirm scope in one line, then proceed without waiting on anything else:\nticker and listing, window, benchmark (`references/data-pulls.md` section 4: the broad index of\nthe stock's own exchange, in its currency), and whether an optional sector factor is wanted.\n\n## Pipeline\n\nWork in a scratch directory. Scripts are in `scripts/`, the page template in `assets/`.\n\n**1. Plan and pull prices.** `python3 scripts/plan.py <ticker> <benchmark> <start> <end>` prints\none Samaya question per slice. Send each as its own `financial_qa` call, poll until done, and\nsave each reply verbatim to `raw/<prefix>_<n>.txt` with the `# url=... title=...` header. Append\nevery Samaya URL returned, from any call, to `urls_seen.txt`. Read `references/data-pulls.md`\nbefore the first pull: a single pull silently keeps only ~100 sessions.\n\n**2. Check integrity.** `python3 scripts/integrity.py <stock> <bench> raw clean`. It drops copied\nbars, flags corrupt volume, logs vendor gaps and repairs a whole-day date shift only when one\nshift is unambiguous. If it blocks, stop and tell the person what failed; never hand-edit data.\nState any repair or gap on the page.\n\n**3. Screen.** `python3 scripts/compute.py <stock> <bench> <start> <end> clean out.json clean=1`\n(add `sector=`, `exclude=` options only per `data-pulls.md` section 6). It grades every day on a\nsignal ladder, from the stock-specific move (z = move beyond what beta x benchmark explains,\ndivided by the stock's normal daily variation) and volume:\n\n| Signal | Rule | Treatment |\n|---|---|---|\n| `strong` | z >= 2.5 on volume >= 1.75x the 50-day median, not an index-mechanics day | Callout, researched |\n| `price` | z >= 3.0 on any volume (mega-caps rarely spike volume on news) | Smaller callout, \"on normal volume\", researched |\n| `rank` | Among the window's largest stock-specific moves, about 10 per year scaled to the window (z >= 2) | Callout, researched (top-move guarantee) |\n| volume-only | Volume >= 2x with z < 1.5 | Listed, never researched or storied |\n| `run` | 10-session cumulative z >= 2.5 with no flagged day inside | Hatched, clickable band with its own panel and legend entry, researched as one window |\n\nVolume on quarterly-expiry and Russell-reconstitution days cannot make a day `strong`. There is\nno cap.\n\n**4. Research every flagged day, run and drift phase.** Follow `references/attribution.md`. Go\ndown the source ladder and stop at the first rung that explains the move: company releases and\nfilings, then news asked as \"Why did <company> (<ticker>) stock move <x%> on <date>?\", then Web\nresults that report an event (not the move) from an attributable publisher, never a \"why it\nmoved\" explainer, then X posts dated that session as a last resort (tier `social`: context, not a confirmed cause).\nNever use Reddit: its posts carry no date. One credible original report with no second source\nis tier `reported` with `single_source: true` (shown as \"single source\"); a relayed copy never\ncounts as a second source. Any phase whose unflagged sessions add up to more than an 8%\nstock-specific move must be researched as a window (`context_sources`); the render blocks\notherwise. Apply the timing rule (a release after the close\nmoves the next session). Measure beat or miss against the company's own guidance from\ntranscripts or 8-Ks, never consensus. Treat rating changes as same-day reactions, in counts.\nAlso collect results releases with times, and context events (results days, management\nchanges, large M&A) that did not clear the screen.\n\nEverything flagged is researched; the page decides how loudly to show it. Numbered markers go to\nthe moves that account for 75% of flagged stock-specific movement, plus every results reaction and\nthe three largest moves; the rest are small numbered circles on the price line with the same panel. Markers\ncluster only when they would overflow the screen's row budget, and never the results or top three.\n\n**5. Write `events.json`.** Schema and limits in `references/events-schema.md`; a complete\nworked example is `references/example-events-aapl.json`.\n\n**6. Find the story, then write it.** `python3 scripts/phases.py out.json events.json` lists the\nallowed phase boundaries; add boundaries to print each phase's return, market and\nstock-specific parts, and flagged-day versus other-session contributions. Write the headline\nand phase tiles from those numbers. Say plainly when the evidence cuts against the obvious\nnarrative.\n\n**7. Render.** `python3 scripts/render.py out.json events.json urls_seen.txt\nassets/template.html <out>.html`. It writes nothing if any gate fails (list in\n`events-schema.md` section 4). Fix the data or the wording, never the validator.\n\n**8. Audit, check and deliver.** Run `python3 scripts/coverage_check.py out.json events.json`: it checks the page against the raw data (the window's largest raw and stock-specific moves covered, as many as the scaled top-move guarantee, no unresearched drift phase) and must report PASS or have each issue explained.\n\n**Then check and deliver.** If a headless browser is available, open the page at desktop and\nmobile widths, confirm no script errors, and look at it. Deliver the single HTML file through\nthe environment's publishing or file mechanism; it loads its fonts from Google Fonts and runs\nnothing else externally.\n\n## Rules the page must keep\n\n- Only Samaya-retrieved information. No figure, date, officeholder or cause from memory.\n- No investment recommendation, price target, valuation view or forecast in any text.\n- Beat or miss only against the company's own prior guidance; say so when no numeric guide\n  exists.\n- No broker or analyst names in visible text; rating changes appear as counts, names stay in\n  source titles.\n- Every source names the actual document or feed and links the exact Samaya URL returned;\n  Samaya is the routing layer, not the source.\n- Unexplained large moves are shown, labelled \"No catalyst found\", not hidden or explained away.\n- X posts are the last source and never a confirmed cause; Reddit is never used.\n- A phase that rose or fell on unflagged days is researched as a window, never described as\n  \"quiet days\" by default.\n- Market-driven large moves get no story; they are listed with their split.\n- Name the benchmark everywhere the split is shown.\n- No model or vendor name as an author anywhere in the output.\n\n## Reference files\n\n| File | Read when |\n|---|---|\n| `references/data-pulls.md` | Before the first candle pull; when integrity blocks; for non-US listings |\n| `references/attribution.md` | Before researching flagged days |\n| `references/events-schema.md` | Writing `events.json` and the story; fixing a render BLOCK |\n| `references/example-events-aapl.json` | As the template for `events.json` |\n"
}

SHA-256 of public snapshot: 187ced8c1543f58901b1bae9e634fe1b30ac1b5a1af6cc090159eb678a60c9dd