skills/cass-context/SKILL.md
Use when the user says: "what context do I have", "relevant history for this task", "what have we done related to this". Also usable proactively when starting work that likely has prior history. Assembles a structured context brief from past session history via cass, scoped to the current task and workspace.
npx skillsauth add NodeJSmith/Claudefiles cass-contextInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Assembles a structured context brief from past session history, scoped to the current task and workspace, backed by cass search --robot. Unlike /cass-recall (open-ended search, freeform synthesis), this skill takes a task description and produces a brief with a fixed shape — designed so cm's cm context CLI could be swapped in later as the retrieval backend without changing what callers get back.
$ARGUMENTS — the task description, e.g. /cass-context "implementing rate limiting for the API". If empty, ask what task the user is about to start.
Extract content-bearing keywords from the task description — specific nouns, technologies, module or file names, domain terms. Drop generic verbs and vague nouns (same filtering as /cass-recall).
Search, scoped to this workspace and recent history:
cass search "<keywords>" --robot --workspace "$(pwd)" --days 30 --limit 5
Read into hits when the snippet isn't enough. Each hit's source_path is a real transcript file — read it directly for more detail than the snippet carries.
Synthesize into the four-section brief below.
Always produce exactly these four sections, in this order, even when a section has nothing to report (write "Nothing found." rather than omitting the section). This contract is the swap point for a future cm context backend — downstream consumers depend on the shape staying stable, not on cass being the source.
### Relevant History
[What was done before related to this task, with dates and outcomes. Cite source_path or session context when useful.]
### Decisions Made
[Past decisions that constrain or inform the current task — architecture choices, rejected alternatives, agreed conventions.]
### Patterns to Follow
[Conventions established in prior work relevant to this task — naming, structure, testing approach, error handling style.]
### Suggested Follow-up Queries
[2-3 `cass search` commands for deeper investigation if this brief isn't enough, e.g.:
- `cass search "<related term>" --robot --workspace "$(pwd)"`
- `cass search "<specific decision keyword>" --robot --days 90`]
Keep the brief tight — this is a working document to orient before starting a task, not an exhaustive history. If the search returns nothing relevant, say so plainly in each section rather than padding with generic advice.
development
Use when the user says: "document how X works", "write up how this works", "durable explanation", "explain this for the docs", "document this subsystem". Writes a durable, architectural-altitude explanation that survives code churn.
development
Use when picking up a fresh session after /clear, a stop, or an unanswered AskUserQuestion. Reconstructs the prior session's intent from its transcript tail and surfaces any unresolved decision; user-invoked only — for a hand-written end-of-day handoff use /mine-good-morning instead.
development
Use when the user says: "what did we discuss", "continue where we left off", "remember when", "as I mentioned", "you suggested", "we decided", "search my conversations", "find the conversation where", "what did we work on", or uses implicit signals like past-tense references, possessives without context, or assumptive questions. Direct search over past Claude Code sessions via cass.
development
Use when the user says: "what would a v2 look like", "how would we rebuild this", "next iteration of this design", "what improvements are we skipping", "what would a mature version look like", "what are we not considering here", "how would we make this more robust", "sophistication ceiling", "elevate this subsystem". Surfaces upward improvements to an existing subsystem through three generator lenses — each candidate annotated with cost and the case against, ordered by signal. A counterweight to solo-dev myopia: it proposes, it never decides.