Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+22 moreClaude CodeCursorCodex CLIMuseDots
OpenClaw
Manus
Grok
Grok Bot+18 agentsContext Compass: Repo Prep & Proof-Driven Changes
Prepare a repo with traceable evidence, then guide requested changes from explicit behavior to test-first proof.
$6
Context Compass: Repo Prep & Proof-Driven Changes
Example session with this skill installed
Prepare a compact baseline for this repository before a follow-up change. Follow its governing instructions. Trace only documented behavior; specifically trace the acceptance criterion for API/schema candidates and behavior-test leads. Inspect source and test assertions, and mark each link Verified, Partial, or Unknown with file evidence. List available quality commands separately from commands run. Do not edit files or run commands. Use inferences only to decide what to inspect; leave unsupported behavior unknown. Return the report in chat as text; do not create a report file.
- Read your context and instructions
- Compiled the context compass
Baseline scope
Verified: docs/acceptance-criteria.md:9-10 documents discovery of common API/schema candidates, likely behavior-test leads, and the caveat that no filename match does not prove there is no contract.
Evidence trace
| Link | Evidence inspected | Status |
|---|---|---|
| Documented expectation | docs/acceptance-criteria.md:9-10 states the candidate-discovery behavior and absence caveat. | Verified |
| Implementation | src/context-compass.ts:211-221 renders contract candidates, test-file candidates, and completeness caveats. src/context-compass.ts:516-529 recognizes listed common filenames and contract/schema directories. | Verified for the documented filename/path behavior; this does not identify which contract governs a real code path. |
| Test assertions | tests/context-compass.test.ts:59-82 asserts the synthetic contract paths, tests/api.test.ts lead, and no-candidate caveat. | Verified by assertion inspection; this baseline did not execute the tests. |
| Real-world coverage | The fixture proves candidate discovery for its named files. The actual contract-to-caller-to-assertion trace for a future behavior is not established here. | Unknown |
Available verification
- Available: npm run check (package.json, check script).
- Run for this read-only baseline: none. No quality command was executed.
Next evidence step
For the next requested behavior, inspect its actual contract, implementation, callers, and test assertions. Do not use a filename match as proof that behavior is covered. If required behavior has no user direction or repository evidence, leave it Unknown and ask before choosing.
Output format: This is a text report in chat. The skill does not create or write a PDF or separate report file.
Connects securely to your tools. The creator never sees your data.
About this skill
Value
Before changing code in an unfamiliar repository, see which requested or documented behavior is traceable—and where proof is missing. Context Compass maps instructions, interfaces, code, callers, tests, docs, and available checks. Its repository baseline marks each trace Verified, Partial, or Unknown. Its change workflow carries one requested behavior from explicit scope through contract inspection, a failing behavior test, implementation, documentation, and verification.
How it works
- Use the bundled scanner as a local index, then inspect actual source and test assertions.
- Keep candidate paths, inferences, and unknowns distinct from verified behavior.
- Separate commands available from commands actually run.
- Keep analysis read-only. Make changes only when requested and allowed by repository instructions.
- If a required behavior decision lacks user direction or direct repository evidence, stop and ask; do not choose by convention.
Where it helps
Prepare an inherited or unfamiliar repo, plan a cross-cutting change, or give a coding-agent handoff a clear evidence trail.
Included and limits
The paid ZIP includes both workflows and a bundled Node.js scanner. The public Context Compass plugin remains free and MIT licensed. The scanner uses Node.js built-ins, makes no network calls, skips symlinks and known private-memory directories, and does not run project commands by default. File-name matches are incomplete leads; the workflow cannot prove every contract or test, guarantee correctness, or issue a whole-repository ready/not-ready verdict. No time, token, or defect savings have been measured. The assistant host may process repository content. The skill returns text; Agensi may show a platform-generated PDF-style card that the skill does not create.
Launch offer: $6.00 USD one-time (at least 30% off) until 2026-11-01T08:51:08.960000Z (01 November 2026 at 09:51:08 CET, Europe/Amsterdam). The regular one-time price is $9.00 USD afterward.
How to use
Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scanned
Verified clean 9 days ago
- Passed all security checks, Safe to install