Skip to content

tough-tongue/toughtongue-skills / skills/ttai-session-analyst

ttai-session-analyst

Verified

Turns 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).

FileKindSize
SKILL.mdskill md9.9 KB
agents/openai.yamlconfig494 B
references/data-model.mdmarkdown6.7 KB
references/report-templates.mdmarkdown4.7 KB
references/reprocess.mdmarkdown2.5 KB

references/reprocess.md

View on GitHub
# Reprocess

One write path: `ttai:analyze_session`. The MCP server marks it as reaching
outside the account (it can fire the Scenario's webhooks). Confirm it with the
user before calling.

## Contents

- Confirm before acting
- Fill missing analysis: `ttai:analyze_session`
- Poll for results
- External calls

## Confirm before acting

State the exact target and wait for an explicit yes: the session IDs, their
Scenario, and what is missing. Example: "Run analysis on 7 sessions of Scenario
`<id>` that have no evaluation? This can fire the Scenario's post-session
webhooks."

Never infer consent from an earlier, broader request such as "build me a
report".

## Fill missing analysis: `ttai:analyze_session`

Argument: `session_id` only.

- The Scenario's analysis settings decide what runs: evaluation, variable
  extraction, and Scenario learning. If none is configured, the call does
  nothing and still returns OK.
- It fills only what is missing. A session that already has `evaluation_results`
  is not re-scored. It cannot regrade under a changed rubric.
- It runs in the background. An OK response only means the attempt was queued.
  If processing is already `running`, the attempt is skipped.
- Use it for sessions with no `evaluation_results`, or after
  `post_session_status.state` is `failed`.
- Do not call it while the state is `running`, and do not loop it.
- A 429 "Analysis limit reached for this session." is final: report it. Any
  other 429 is a rate limit: wait, then retry.

## Poll for results

Poll with `ttai:list_resources(sessions)` (`ids`,
`include_fields: ["evaluation", "processing"]`) or `ttai:get_resource(sessions)`. Back off
between reads, starting at about 15 seconds, and stop after a few minutes.

- **`state: running`** — in progress. Keep polling.
- **`state: failed`** — the last run errored (`failed_reason`). Report it. Retry
  once with `ttai:analyze_session` only if the user approves.
- **`state: idle` and `evaluation_results` present** — done. Build the report.
- **`state: idle`, no `evaluation_results` after a few polls** — not configured,
  or nothing ran. Tell the user the Scenario may not have analysis enabled. Do
  not retry in a loop.

## External calls

No MCP tool ingests an external call transcript or recording. Scoring one is a
REST API call that needs an API token set up outside this skill: send the user
to [Create session](https://app.toughtongueai.com/docs/api-reference/sessions/create).
Once it exists, its session reads and reprocesses like any other.
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: 472kUpdated Claude CodeCodexVerifiedInstall
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: 323kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 323kUpdated Claude CodeCodexVerifiedInstall
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 availableStars: 323kUpdated Claude CodeCodexVerifiedInstall