Fair Price Brief

Fair-Price Negotiator's Brief

You have a live quote from a creator (or their manager) and you need to know: is it fair, and what do I say back? This skill profiles the creator, benchmarks the quote against the Creatorland deal corpus for the matching vertical and deal type, and ends in a one-page negotiation memo a brand-side manager or agency negotiator can walk into the call with. The clearest ROI in the catalog: one avoided overpay covers a year of pro.

Read first: ${CLAUDE_PLUGIN_ROOT}/shared/conventions.md (tool schemas, credit prices, the seven conventions). This skill honors thrifty/thorough credit modes (${CLAUDE_PLUGIN_ROOT}/shared/credit-modes.md), Refusal Recovery (${CLAUDE_PLUGIN_ROOT}/shared/refusal-recovery.md), and the Freshness Gate (${CLAUDE_PLUGIN_ROOT}/shared/freshness-gate.md).

Calibration: if ~/.claude/plugins/config/creatorland/creatorland-data/PROFILE.md exists (written by /intake), read it first — persona, familiarity (1–3), and saved defaults set narration level and assumed context (conventions.md §15).

Inputs to collect

• Creator identifier (required) — handle, profile URL, or email. Map to one of the five identifier types: a handle/URL → social_handle, an email → email. Never ask for an identifier format the user already gave you in another form.
• Quoted fee (required) — the number on the table, in USD. If quoted in another currency, convert and note the conversion in the memo.
• Deal composition (required) — what the fee buys, e.g. "1 IG reel + 3 stories". Map to the closest deal_type the corpus supports (e.g. "sponsored_post", "affiliate"); if the composition is a bundle, benchmark the dominant deliverable and say so in the memo.
• Brand / vertical context (ask only if not inferable) — if the creator's profile makes the vertical obvious (a skincare creator quoted by a beauty brand), don't ask. Ask when the profile is multi-vertical or the brand's category differs from the creator's content.
• Negotiation posture (optional, default "find middle") — one of: hold firm (counter low, anchor hard), find middle (counter toward median), walk-away threshold (user names a ceiling; memo says whether the band supports it). Shapes the counter range and the tone of talking points.

Never ask twice; never ask for things the conversation already contains.

Flow

The server-side fair-price-check prompt exists and is maintained with the server — use it as the analytical spine for step 4's positioning logic where the harness exposes server prompts.

1. Profile the creator — get_creator_profile { identifier: { type: "social_handle", platform: <platform>, handle: <handle> } } (or { type: "email", email: ... } etc., exactly one type). From the response take: interests + brand affiliations (pro) → infer the primary vertical; follower count → tier framing; audience geo; data freshness (Freshness Gate: if the profile is stale, the memo must carry a "re-verify before negotiating" flag — drifted follower counts change the tier framing). 1 credit.

2. Benchmark the quote — query_market_intelligence { mode: "rate", vertical: <inferred vertical>, deal_type: <mapped deal type>, quoted_rate: <fee in USD>, creator_tier: <if the creator's follower tier is known> } When the creator's follower size is known, add creator_tier (emerging <1k / nano 1k-10k / micro 10k-100k / mid 100k-500k / macro 500k-1M / mega 1M+) so the counter is benchmarked against same-size creators in the vertical — this is the stronger comp. It requires a vertical; a tier too thin for the distinct-creator privacy floor broadens (disclosed) to the all-tier band, so note the clearance level in the memo's provenance line. — wrapped in Refusal Recovery. Rate mode enforces a privacy floor of 10 brands / 50 deals; if the narrow slice refuses, walk the ladder (drop deal_type → drop sub_category → drop active_since → drop company_type → vertical only), stop at the first rung that clears, and disclose the clearance level in the memo ("benchmark holds at the Beauty vertical level — the 'Beauty × sponsored_post' slice was below the privacy floor"). Thrifty mode: at most 2 rungs, then report what you have. 5 credits per call, including each ladder retry.

3. Comp set (optional — thorough mode only) — search_creators { mode: "lookalike", seed_creator: { type: "social_handle", platform: <platform>, handle: <handle> }, limit: 5 } — names comparable creators to use as negotiation anchors ("comparable talent in this band includes…"). Inference-free, 2 credits. Skip in thrifty mode. Do NOT profile the comp set (that's a 5-credit fan-out for names that only need to exist); if the user wants comp-set detail, that's a follow-up run.

4. Compose the memo (no tool calls) — position the quote within the band ("the quote sits at ~pXX"), derive the counter range from the posture, and write three talking points that each cite a specific number from steps 1–2. The credit-modes pre-run estimate fires only if a run will exceed ~30 credits (rare here — only a deep refusal ladder gets close).

Deliverable

A one-page negotiation memo in markdown, paste-ready:

# Negotiation brief — @<handle> · <deal composition>

**Quote on the table:** $<fee> for <composition>
**Market band (<benchmark scope>):** p25 $<x> · median $<y> · p75 $<z>
**Position:** the quote sits at ~p<XX> of the band — <above/within/below
the typical range for this vertical and deal type>.

## Suggested counter
$<low>–$<high> (<posture rationale: e.g. "anchors at median with room to
settle at p60, consistent with 'find middle'">)
<If walk-away threshold supplied: "Your $<ceiling> ceiling sits at ~p<XX>
of the band — the data <supports/does not support> holding it.">

## Talking points
1. <Data-grounded point citing a band number>
2. <Point grounded in deal volume / recency signals from the rate response>
3. <Point grounded in the creator's tier/audience-geo fit vs the band>

## Comparable talent (negotiation anchors)        [thorough mode]
@<comp1>, @<comp2>, @<comp3>, @<comp4>, @<comp5> — lookalikes of the
quoted creator; same vertical band applies.

---
**Benchmark basis:** <vertical / scope at which the privacy floor cleared,
with the floor-disclosure note if the query was broadened — "this is a
privacy feature of the data source, not missing data">
**Provenance:** <the provenance line the rate response returned, verbatim>
· recency window: <window>
**Profile freshness:** <fresh/aging/stale per the Freshness Gate — if stale:
"re-verify follower tier before negotiating">
**Honesty note:** the band is the corpus-level market range for this
vertical/tier — it is NOT this creator's personal rate history (that data
does not exist).
Credits used this run: ~N (breakdown: 1 profile + 5×<rate calls> + 2 comp set)

Never strip the provenance or scope footers to make the memo prettier — they are what makes it defensible when pasted into a client thread.

Honesty rules

• Market band, never "their rate." The benchmark is corpus-level for the vertical/tier. Per-creator rate history does not exist in the product; never phrase the band as what this creator "usually charges" or "has accepted."
• Disclose every broaden. If Refusal Recovery widened the slice, the memo states the level at which the floor cleared. Never present a broadened number as if it answered the narrow question.
• Bundles are approximations. When the deal composition is a multi-asset bundle benchmarked against a single deal_type, say so.
• Stale profile = flagged memo. A stale profile changes tier framing; the Freshness Gate note is not optional.
• No contact info. Never imply you can reach the creator; route contact questions to Creatorland connections (coming soon) or public profiles.

Credit footprint

thorough: ~8–13 credits (1 profile + 1–2 rate calls + comp set; deep refusal ladders add 5/rung) · thrifty: ~6 credits (1 profile + 1 rate call, max 2 ladder rungs, no comp set)