session

Plugin session from marketplace skills in tkstang/skills. Commands read from the repo's manifest — 6 days ago.

Skills (14)

• align

  Use when the user explicitly asks to align on their goals and underlying problems, at the start of work or when an ongoing session or project feels off course.

• debrief

  Use when the user returns to a session and wants a concise account of what happened while they were away, including what changed, what was decided, what is shown versus only reported, and what is waiting on them. Looks back; it does not recommend or execute next actions.

• explain-options

  Use when the agent offered choices, just now or earlier (a question, a list, "A or B", or an AskUserQuestion) and the user wants context to decide: "explain the options", "what are the trade-offs", "why pick one over the other", how they differ, or more detail on them. Explains background, trade-off

• export-transcript

  Use when the user asks to export, save, or download the current coding-agent conversation as a Markdown file (e.g. "export this session transcript", "save the conversation as markdown"). Locates the live transcript via an announced session marker, drops tool calls and hidden injected payloads, and w

• fork-to-destination

  Use when the user wants to continue a Claude Code or Codex session in an existing destination worktree, including importing supported history into the other provider before running its native fork command.

• handoff

  Use when the user asks to hand off, transfer, or let another agent continue coding work. Produce concise evidence-grounded continuation context with an optional accepted sanitized transcript link.

• messaging

  Use when three or more local coding-agent sessions need addressed questions, blockers, review requests, or handoffs without sharing transcripts. Provides a durable manual inbox and collaboration log.

• must-we

  Use when the user questions whether a claimed blocker, requested approval, review requirement, or proposed extra work is actually necessary, and wants an evidence-based recommendation and a simpler alternative when warranted.

• next-steps

  Use when the user wants a clear recommendation for what to do next based on the conversation, with enough context and reasoning to understand the actions and who needs to take them.

• reorient

  Use when the user returns to a long thread after time away, or the agent loses the thread after compaction or a change of direction, and the goal, the side quests, and where things stand need rebuilding. Reads the whole session transcript, checks the work's current git and PR state read-only, and en

• retro

  Use when the user asks to review a recently used skill or a bounded coding session to learn what should improve next time. Produces evidence-backed findings and proposed validation without editing skills, rules, memory, installs, or publications.

• search

  Use when the user wants to find a past coding-agent session or conversation (e.g. "find the session where we…", "search my sessions for…", "which conversation did we discuss X in?"). Searches local Claude Code, Codex, and Cursor history, metadata, and transcripts from agent-expanded patterns plus op

• seshtk-cli

  Use when the user asks to use or understand seshtk, the Session Toolkit CLI, for local coding-agent session identity, discovery, search, transcript reads, Markdown export, activity capture, or Codex evidence. Also use for CLI capability checks and troubleshooting. Keep explanation-only questions ind

• sitrep

  Use when the user asks where the current work stands against its goal or plan, including what is complete, in progress, remaining, at risk, or awaiting a decision. Reports progress across the established scope; it does not recap an interval or recommend next actions.

Similar plugins