# Change Log

## v1.8.1 - 2026-04-23

- polished the browser UI to make Wireshark menu paths render as explicit navigation cues instead of plain sentence text
- expanded menu-path emphasis across assistant guidance, handrail instructions, alternate routes, and reference captions/titles
- improved guided-card readability so beginner users can follow “where to click next” more easily

## v1.8.0 - 2026-04-23

- added the `v1.8` guided handrail system with metadata-driven next-step selection across the built-in playbooks
- expanded first-class Wireshark workflow guidance with reference assets for Conversations, Endpoints, Protocol Hierarchy, Expert Information, Follow Stream, IO Graphs, Flow Graph, Packet Lengths, and TCP RTT
- made investigation progression linear for transcript/export consistency and improved the investigation trail UI
- strengthened protocol-aware guidance for TCP, DNS, HTTP, suspicious traffic, and Wi-Fi / 802.11, while marking BLE / BTLE coverage as limited
- added lightweight baseline save/compare support and protocol-native payload expansion in the receiver and Lua handoff

## v1.7.2 - 2026-04-21

- expanded the built-in playbook set with `DNS Investigation` and `Web / HTTP Issue`
- strengthened the TCP, suspicious-traffic, Wi-Fi, and BTLE playbooks using practical packet-analysis workflows
- improved playbook recommendation logic so DNS and HTTP packet context can suggest more relevant guided investigations
- added regression coverage for the new DNS and web playbook paths

## v1.7.1 - 2026-04-21

- cleaned up chat bubble spacing so labels and message text no longer run together
- changed user message headers from pill labels to plain `YOU:` text for clearer attribution
- switched assistant and status labels to flat pills without drop shadows
- made assistant titles render on a new line with bold headings for playbook and next-step messages

## v1.7.0 - 2026-04-21

- promoted the codebase to the `v1.7.0` release line
- strengthened beginner and SOC L1 triage guidance across playbooks and generic packet-analysis flow
- improved protocol-aware scoping for Wi-Fi and BTLE device-focused filters
- added AI coaching prompts so AI next-step guidance teaches workflow instead of only returning an answer
- aligned built-in AI prompt profiles with the training-aid approach
- completed a gold-master decision test pass and documented release strengths and weaknesses in `docs/v1.7-gold-master-decision-test.md`

## v1.6.0 - 2026-04-20

- added a reusable `systemd` service template under `deploy/systemd/sharkbot.service`
- reorganized the README into setup, run, usage, systemd, and change-log sections
- changed the receiver startup so debug mode is off by default and only enabled with `--debug`
- added in-memory investigation continuation so Wireshark can push new packet context into an existing chat session
- added a downloadable Markdown investigation export from the browser UI
- moved investigation ID and export controls into a persistent bottom-left sidebar usage section
- added specialist AI prompt profiles:
  - `specialist`
  - `packet_analyst`
  - `incident_response`
- improved AI provider routing so normal filter requests stay rule-based unless AI is explicitly requested
- added Anthropic overload handling with retry/backoff for HTTP `529`
- improved missing-session handling so stale browser reloads show a friendly error page instead of a traceback
- improved remote/headless receiver support with separate bind and public browser URLs
- refined the UI onboarding flow, AI reconfiguration flow, packet card styling, response labeling, and filter highlighting