code-forensics-stinger skill
scriptsPerforms forensic reviews of software vendor engagements by analyzing contracts, invoices, communications, and git logs to produce cited evidence drafts.
- GitHub stars
- 95
- Files
- 40
- Last commit
- Oct 18, 2026
- License
- AGPL-3.0-or-later
- Allowed tools
- —
What does the code-forensics-stinger skill do?
code-forensics-stinger performs forensic reviews of software vendor engagements by analyzing contracts, invoices, communications, and git logs to produce cited evidence drafts.
Summary generated from the skill's SKILL.md.
About the code-forensics-stinger skill
Performing forensic reviews of software vendor engagements is the primary function of this skill. It assists users in analyzing contracts, invoices, communications, and git logs to produce cited evidence drafts intended for human or legal counsel review. An agent should load this skill when a user provides source records and requests a formal investigation into an agency or software engagement.
The workflow moves through several phases including email processing, invoice extrapolation, git log forensics, and audit log analysis. It produces structured outputs such as attorney memos, agency reports, or pre-litigation packs using specific scripts like build_master_report.js or build_attorney_memo.js. The system relies on a case-facts.json file which contains matter metadata and an evidence register.
This tool focuses on documenting observed activity and billing estimates while maintaining a strict separation between source records and validated inputs. It does not provide legal advice or automatic external sends. Users must provide the case-facts file and specify an output directory for the generated docx or markdown files.
What an agent can do with it
- 1Analyze vendor billing discrepanciesReview invoices and communication records to document billing and create separate estimates.
- 2Investigate software delivery historyExamine git logs and audit logs to identify observed activity and coverage gaps.
- 3Prepare legal review documentsGenerate a combined draft for counsel review using the pre-litigation rendering option.
- 4Verify software dependency securityResearch version evidence and dated advisories for software dependencies.
Written from the skill's SKILL.md. Every use case is checked against a quote from it.
How to install the code-forensics-stinger skill
Each command copies the skill folder into the client's skills directory. Restart the client or start a new session to load it.Install the code-forensics-stinger skill in Plugin
/plugin marketplace add legioncodeinc/vibe-coding-tools /plugin install code-audit-tools@wasp-nest
Install the code-forensics-stinger skill in Claude Code
~/.claude/skills/code-forensics-stinger/SKILL.md
$ npx degit legioncodeinc/vibe-coding-tools/plugins/code-audit-tools/skills/code-forensics-stinger ~/.claude/skills/code-forensics-stingerInstall the code-forensics-stinger skill in Claude Code (project)
.claude/skills/code-forensics-stinger/SKILL.md
$ npx degit legioncodeinc/vibe-coding-tools/plugins/code-audit-tools/skills/code-forensics-stinger .claude/skills/code-forensics-stingerInstall the code-forensics-stinger skill in Codex
~/.agents/skills/code-forensics-stinger/SKILL.md
$ npx degit legioncodeinc/vibe-coding-tools/plugins/code-audit-tools/skills/code-forensics-stinger ~/.agents/skills/code-forensics-stingerInstall the code-forensics-stinger skill in Gemini CLI
~/.gemini/skills/code-forensics-stinger/SKILL.md
$ npx degit legioncodeinc/vibe-coding-tools/plugins/code-audit-tools/skills/code-forensics-stinger ~/.gemini/skills/code-forensics-stingerInstall the code-forensics-stinger skill in Claude.ai
npx degit legioncodeinc/vibe-coding-tools/plugins/code-audit-tools/skills/code-forensics-stinger code-forensics-stinger && zip -r code-forensics-stinger.zip code-forensics-stinger # claude.ai → Customize → Skills → + → Create skill → Upload a skill → code-forensics-stinger.zip
SKILL.md
View on GitHubCode Forensics Stinger
Use this skill when the user requests a forensic review of a software or agency engagement and can supply source records. Ask who the parties are and what happened. Never recognize a new matter by assuming a named party, country, price, jurisdiction or misconduct pattern from another matter.
Read README.md and principles first. For routine code review, route to the relevant engineering specialist. Legal determinations belong to counsel.
Workflow
| Phase | Guide | Result |
|---|---|---|
| 0 | Intake | Agreed scope, materials and private case workspace |
| 1 | Source-indexed communication records | |
| 2 | Invoices | Documented billing and separate estimates |
| 3 | Git | Acquired history with stated limits |
| 4 | Dependencies | Version evidence and dated advisories |
| 5 | Logs | Observed activity and coverage gaps |
| 6 | Marketing | Contracted outputs and comparable metrics |
| 7 | Synthesis | Author-supplied report sections and cited output |
| 8 | Review letters | Combined draft for counsel review |
Assess each phase for relevance. Record absent materials without fabricating findings. Keep source records and working notes separate from the validated renderer input.
Rendering
The input contract and JSON Schema describe case-facts.json. It contains matter metadata, an evidence register and authored document sections. It is not an automatic investigation accumulator. Python helpers do not populate it automatically.
| Kind | Entry point |
|---|---|
| master | scripts/build_master_report.js |
| agency | scripts/build_agency_report.js |
| attorney | scripts/build_attorney_memo.js |
| plain_language | scripts/build_plain_language.js |
| pre_litigation | scripts/build_pre_litigation.js |
Every invocation requires --case-facts FILE --out-dir DIR. Default format is docx; use --format markdown for Markdown. Missing input or a missing selected document refuses before any output directory is created. Existing output files are never overwritten. Generate revisions in a new private directory.
Review standard
Every material claim has source IDs and an evidence tier. Estimates are separate from observed facts. Do not promise a fixed page count, number of letters, spreadsheet tabs, outcome, recovery amount or litigation readiness. The renderer copies supplied text; review it before sharing. Public training fixtures are explicitly fictional and must never be used as a real case's facts.
Critical Directive
Read relevant guides and inspect the actual source before relying on a tool. Supplement missing research with current authoritative sources. Preserve originals unmodified. No legal advice, assumed privilege, automatic external sends or unapproved live actions. Security and Quality review apply before any authorized commit; the controlling user must explicitly authorize commit, push, publication and distribution.
Files in this skill
40code-forensics-stinger skill questions
How do I install the code-forensics-stinger skill in Claude Code?
Run npx degit legioncodeinc/vibe-coding-tools/plugins/code-audit-tools/skills/code-forensics-stinger ~/.claude/skills/code-forensics-stinger. Claude Code loads skills from ~/.claude/skills (all projects) or .claude/skills (one project) and uses a skill when its description matches the task.
Which clients can use the code-forensics-stinger skill?
Any client that reads the Agent Skills format: Claude Code, Codex, Gemini CLI and claude.ai (upload as a ZIP). Scripts in the skill run only where the client allows code execution.
What types of documents can this skill analyze?
The skill reviews software vendor engagements using contracts, invoices, messages, delivery records, git logs, and audit logs.
Can the tool automatically generate a full investigation report?
No, the case-facts.json file is not an automatic investigation accumulator and Python helpers do not populate it automatically.
What file formats can the reports be generated in?
The default output format is docx, but you can specify markdown using the --format flag.