Update to Zime

{
  "description": "Shows what customers actually say about competitors. Where they come up, what they claim, and how we compare. Only uses real quotes from real calls.",
  "included_files": [],
  "name": "competitive-intelligence",
  "skill_md_contents": "---\nname: competitive-intelligence\ndescription: \"Shows what customers actually say about competitors. Where they come up, what they claim, and how we compare. Only uses real quotes from real calls.\"\nlicense: MIT\nmetadata:\n  zime:skill-id: 97b63479-93dc-4cf2-9ee1-754a216267b4\n  zime:tag: intelligence\n  zime:roles: ae,se\n  zime:visibility: global\n  zime:skill-version: \"1\"\n  zime:marketplace-version: 0.1.20\n---\n\n# Competitive Intelligence\n\nBefore anything else, call the Zime connector tool `skill_used` with skill_id `97b63479-93dc-4cf2-9ee1-754a216267b4`, skill_version `1` and skill_slug `competitive-intelligence`. It records which skill ran for Zime usage analytics. If the Zime connector is not available, skip this step and continue.\n\nAnswers \"what are customers actually saying about Competitor X?\" from the\nworkspace's own calls. The value is that it's evidence from real\nconversations, not what the internet says about a competitor — and the two are\noften very different.\n\n## The line this skill holds\n\nGeneral market knowledge about a competitor is the failure mode. What their\nwebsite claims, their published pricing, their reputation — none of that is\nwhat this skill returns. If the corpus has nothing on a competitor, the answer\nis \"nothing in our calls mentions them\", not a summary from training data.\nThat distinction is the whole point: a rep can act on \"three customers raised\ntheir SSO gap last quarter\" and cannot act on \"they're generally seen as\ncheaper.\"\n\n## How It Works\n\n```\n┌─────────────────────────────────────────────────────────────────┐\n│                  COMPETITIVE INTELLIGENCE                        │\n├─────────────────────────────────────────────────────────────────┤\n│  SCOPE FIRST (this skill decides which)                          │\n│  ✓ \"on the Acme call\"    → list_meetings, then name the call in the │\n│                             ask_zime question                │\n│  ✓ \"across our deals\"    → ask_zime directly (no resolve needed)  │\n├─────────────────────────────────────────────────────────────────┤\n│  DELEGATE (Zime agent)                                           │\n│  + Agent reads real calls, mentions, and extracted signals       │\n│  + Returns evidence with accounts and dates attached             │\n├─────────────────────────────────────────────────────────────────┤\n│  NEVER                                                           │\n│  ✗ Competitor facts from general knowledge                       │\n│  ✗ A \"typical\" objection nobody actually raised                  │\n├─────────────────────────────────────────────────────────────────┤\n│  LOCAL FALLBACK (no zime-mcp)                                    │\n│  ~ Scan a transcript file the user provides                      │\n└─────────────────────────────────────────────────────────────────┘\n```\n\n## Usage\n\n```\n@competitive-intelligence <competitor> [+ scope] [+ when]\n```\n\nIn the ChatGPT desktop app you can also type `/competitive-intelligence`.\n\nReport competitive intel on the competitor the user named.\n\n## Routing\n\n- Objections on ONE deal, competitor or not, or win/loss rates and pipeline\n  totals → `review-account`.\n- A recap of one call that happened to mention a competitor → `review-call`.\n- Turning competitive evidence into a battlecard or talking points →\n  `create-sales-asset`.\n\n## What I Need From You\n\nThe competitor name. Scope and window help: \"on the Acme call\" is a different\nquestion from \"across our deals this quarter\", and this skill routes them\ndifferently.\n\n## MCP mode (required when zime-mcp is connected)\n\n**Required tools:** `ask_zime` (all scopes). For single-call scope also\n`list_meetings` first, to name the exact call in the question.\n\n> `ask_zime` routes to Zime's global agent across the whole accessible\n> corpus, with no separate scoping argument — narrow it by naming the call,\n> account, or window inside the question text. Answering from general\n> knowledge about the competitor while it's available is a failure of this\n> skill.\n\n### Choosing the scope\n\n| Request shape | Path |\n|---|---|\n| \"did Competitor X come up on the Acme call\" | `list_meetings` to confirm the call, then name it in the question |\n| \"where is Competitor X showing up\" | `ask_zime` directly |\n| \"how do we position against them in deals we won\" | `ask_zime` directly |\n| \"what did they say about them last quarter\" | `ask_zime` with the window in the question |\n\nDon't narrow to one call when the question is corpus-wide — that would\nsilently answer a much smaller question than the one asked.\n\n### Cross-corpus\n\n```json\n{ \"question\": \"Where has Initech come up in our calls in the last quarter, what did customers say about them, and how did we position against them?\" }\n```\n\nSend the question close to verbatim with the competitor named explicitly and\nthe time window stated in the text — `ask_zime` has no memory of this\nconversation and no separate date parameters.\n\n### Single call\n\nResolve first, so the call is named correctly:\n\n```json\n{ \"query\": \"Acme\", \"start_date\": \"2026-08-01\", \"end_date\": \"2026-08-13\", \"recorded\": true }\n```\n\nthen name it in the question:\n\n```json\n{ \"question\": \"On the Acme call on Aug 12, did Initech or any competitor come up? What exactly was said, and how did we respond?\" }\n```\n\n### Outcomes\n\n- **Evidence returned** — deliver per Output below.\n- **Nothing found** — say plainly that nothing in the accessible calls\n  mentions this competitor, and stop. Do **not** supplement with what you know\n  about them from elsewhere. A clean \"no mentions\" is a real answer and often\n  a useful one.\n- **An error** — `{\"error\": \"<CODE>\", \"message\": \"<sentence>\"}`. Show the user\n  the message. `INTERNAL_ERROR` retry once; `TIMEOUT` narrow the scope or\n  window and retry; `CREDIT_LIMIT` stop and relay the message; `UNAUTHORIZED` /\n  `FORBIDDEN` means reconnect or lack of access. If it fails twice, say so\n  and offer the local fallback.\n\n## Output\n\nEvidence-first, because a claim without an account and a date can't be used in\nfront of a customer.\n\n```markdown\n**Competitive intel — [competitor]** · [scope] · [window]\n\n### Where they came up\n| Account | What was said | When |\n|---|---|---|\n| [account] | [what the customer actually said] | [date] |\n\n### Recurring patterns\n- **[pattern]** — [accounts it appears in] — [what it means]\n\n### How we're positioned\n[the agent's read, relayed as-is]\n\n### Gaps\n- [dimension the corpus had nothing on]\n\n_From the workspace's own calls via Zime. Not market research._\n```\n\nRules:\n\n- Every row needs an **account and a date**. An undated competitive claim is\n  not usable and shouldn't be presented as intel.\n- Relay what the customer said close to the agent's phrasing — don't sharpen\n  it into a punchier line. \"They said the integration worried them\" is not\n  \"they're losing on integration\".\n- Anything in quotation marks must come from `get_transcript`, not from an\n  agent's paraphrase. Fetch it if the rep needs a literal quote.\n- Keep the Gaps section. Where the corpus is silent is itself competitive\n  intelligence — it usually means reps aren't asking.\n- Never blend in general knowledge about the competitor, even to \"add\n  context\". Label the boundary explicitly.\n\n## Tips\n\n1. **Scope it deliberately** — one call and the whole corpus are different\n   questions with different answers.\n2. **Put the window in the question** for `ask_zime` — it has no date\n   parameters.\n3. **\"No mentions\" is a finding** — it can mean the competitor isn't in play,\n   or that nobody's asking. Both are actionable.\n4. **Need a battlecard?** Pass this evidence to `create-sales-asset`.\n\n## Local mode (only when no zime-mcp server is connected)\n\nIf the user provides transcripts, scan those files only for competitor\nmentions. Open with one line saying the scan covers just the provided files,\nso this is not a corpus-wide view. Same rule: no general knowledge about the\ncompetitor, only what the files contain.\n\n## What this sends where\n\nMCP mode sends the question text, with any call/account/window named inside\nit (to `ask_zime`), and query words plus dates when confirming a call\nfirst (to `list_meetings`). Local mode reads only the provided files. At the\nstart, the skill calls the Zime connector tool `skill_used`, which records the\nskill's id, version and name with your user and session in Zime's usage log.\n\n## Related Skills\n\n- **create-sales-asset** — turn this evidence into a battlecard\n- **review-account** — competitor dynamics on one specific deal\n- **review-call** — the exact quotable line from one call\n"
}