tough-tongue/toughtongue-skills / skills/ttai-session-analyst
ttai-session-analyst
VerifiedTurns Tough Tongue AI sessions into evidence-backed reports. Reads sessions, scores, report cards, strengths, weaknesses, and transcripts from practice roleplays, AI phone calls, and meeting-bot calls for a Scenario, person, or date range, then writes team performance, Scenario health, or individual coaching reports with action items. Can also fill in missing analysis (ttai:analyze_session) after the user confirms. Use when the user asks "how is my team doing?", "top improvement areas for scenario X", "lowest-scoring sessions", "coaching report for Priya", or "session trends this month". Not for changing a Scenario, rubric, or prompt (ttai-agent).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.9 KB |
| agents/openai.yaml | config | 494 B |
| references/data-model.md | markdown | 6.7 KB |
| references/report-templates.md | markdown | 4.7 KB |
| references/reprocess.md | markdown | 2.5 KB |
references/report-templates.md
View on GitHub# Report Templates
Fill every bracket. Drop a section that has no data rather than padding it.
Scores are on the 0–10 scale of `final_score` and `report_card[].score`.
## Contents
- Team performance report
- Scenario health report
- Individual coaching report
- Pattern map (from call transcripts)
## Team Performance Report
```markdown
# [Scenario Name]: Team Performance
**Window**: [from] – [to] · **Sessions**: [N analyzed] ([M] excluded: no
evaluation yet) **Team**: [organization or group], [K] participants
## Score Summary
- Average: [x.x]/10 · Median: [x.x] · Range: [lo]–[hi]
- Trend: [improving / flat / declining] ([weekly averages])
## Skill Breakdown (report-card topics)
| Topic | Avg score | Weight | Reading |
| ------- | --------- | ------ | ------------------------- |
| [topic] | [x.x]/10 | [w]% | [one-line interpretation] |
## Top Improvement Areas
### 1. [Behavior, named concretely]
- Seen in [n]/[N] sessions
- Evidence: "[short transcript quote]" ([participant], [date], [analytics_url])
- Action: [specific practice drill or coaching move]
## Standouts
- [Most improved participant, or best session worth sharing, with analytics_url]
## Recommended Actions
1. [Action, owner, by when]
```
## Scenario Health Report
Use when the question is whether the Scenario itself works. If the verdict is a
Scenario fault, hand the diagnosis and quotes to the `ttai-agent` skill.
```markdown
# Scenario Health: [Scenario Name] ([scenario_id])
**Window**: [from] – [to] · **Sessions**: [N]
## Vital Signs
- Completed: [x]% ([completed] of [total]); terminated: [y]%
- Average duration: [m] min (from legacy `duration`, in seconds)
- Average score: [x.x]/10 · Analysis coverage: [x]% have an evaluation
- Learning flags: [n] sessions with `learning_results.status = issue`
## Symptoms
- **[Signal, e.g. 40% of sessions under 2 min]** — [data]. Suspected cause:
[e.g. the AI ends the call early].
- **[Signal, e.g. one topic scores < 3 for everyone]** — [data]. Suspected
cause: [e.g. the rubric asks for something the roleplay never allows].
## Transcript Evidence
- "[quote showing the failure]" ([analytics_url])
- Learning note: [learning_results.issue / suggestion, if present]
## Verdict
- [Scenario fault → hand to ttai-agent with this diagnosis]
- [User skill gap → coaching recommendation instead]
```
Signals that point at the Scenario rather than the people: most sessions end
early or as `terminated`; one topic scores low for everyone, including strong
performers; transcripts show the AI breaking character, repeating itself, or
ignoring answers; `learning_results` flags the same issue repeatedly.
## Individual Coaching Report
```markdown
# Coaching Report: [Name]
**Window**: [from] – [to] · **Sessions**: [N] on [Scenario(s)]
## Progress
- Score trend: [first] → [latest] ([direction])
- Strongest topic: [topic] ([x.x]/10) · Focus topic: [topic] ([x.x]/10)
## What's Working
- [strength, with an evidence quote]
## Focus Areas (max 3)
### 1. [Concrete behavior]
- What happens: [pattern from weaknesses and transcripts]
- Try this: [from improvement_results.action_items where available]
- Practice: re-run [Scenario name] focusing on [topic]
## Next Check-in
- [Date or session-count target]
```
Keep it to one page. Three focus areas at most; more is a list, not coaching.
## Pattern Map
Use when the user wants a practice Scenario built from real calls ("what do
prospects actually say?"). Pull transcripts for the population, fill this map,
and hand it to the `ttai-agent` skill, which turns it into objections, persona
pools, and rubric gates. A pattern makes the cut when it appears in at least
three sessions; quote the best line heard, marked tone only, never as a script.
```markdown
# [Scenario or call type]: Pattern Map
**Corpus**: [N] sessions, [from] – [to], [who the humans were]
## Stages
| # | Stage | What good looks like (session) | Common miss (session) |
| - | ----- | ------------------------------ | --------------------- |
## Counterpart moves (3+ sessions each)
1. **"[what they say]"** — [n] sessions ([ids]). Beat it by: [the move]. Best
line heard (tone only): "[quote]" ([session])
## Red flags (cap the score when present)
| Pattern | Sessions | Example line |
| ------- | -------- | ------------ |
## Persona pools
- **Level 1 ([n] sessions)**: [who they are, what they ask, what stops them].
Rotate: [2–3 profiles]. Moves pool: [numbers from the list above]
- **Level 2 ([n] sessions)**: [the researched, guarded version]
## Rubric seeds
- Score the human per stage; weight the stages where sessions were won or lost
- Red flags cap the score even when the outcome was good
- Facts the human states must match [source]; an unsourced one costs points
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |