savage-resume

Savage Resume Framework

A Claude Code plugin that runs a savagely self-critiquing, multi-agent review pipeline on a resume, CV, or portfolio and emits a no-bullshit markdown report grounded in the 2026 job market (and where it's heading).

What it does

Point it at a local resume file (.pdf, .md, .txt, .docx) or a portfolio URL. It reads the material, extracts every claim (metrics, titles, companies, skills, differentiation), dispatches five specialist agents in parallel (market-demand, claim-audit, rejection-modes, relevance trajectory, positioning), synthesizes their findings, runs an adversary that tries to kill the synthesis, revises against the kill-list, self-audits against hard invariants, and emits a report to <dir>/.resume-review/.

No fluff. No "you're doing great." If the resume is dead in 2026, it says so and tells you why. If the role is being eaten by AI in 36 months, it says that too, with sources.

Usage

/resume <local-path-or-url>

Examples:

/resume ./my-resume.pdf
/resume ./test-fixtures/bullshit-resume/resume.md
/resume https://example.com/portfolio

Install (Claude Code Plugin)

Add this repo as a plugin marketplace and install, directly from Claude Code:

# Inside Claude Code:
/plugin marketplace add AliBassam/savage-resume
/plugin install savage-resume@savage-resume

That's it. No clone, no per-project setup. The /resume command is available in every Claude Code session on your machine.

To uninstall:

/plugin uninstall savage-resume@savage-resume

To verify it's loaded:

/plugin

savage-resume should appear in the plugins list, with /resume available as a command.

How to tell it worked

After /resume <path-or-url>, check <dir>/.resume-review/ (for a local file, <dir> = the file's parent directory; for a URL, the current working directory):

• The top-level YYYY-MM-DD-HHMM-<subject>.md is the final report.
• runs/YYYY-MM-DD-HHMM/ contains the full working trail: ingest summary, claim list, all five scratchpads, every draft version, kill-list, and audit results.

What the report contains

1. Verdict — dead, savable with surgery, or shipping-ready. One paragraph, no hedging.
2. What it actually is — neutral summary of the subject as written.
3. Claims vs. reality — every metric, title, and impact claim, verified against independent evidence.
4. Market demand (2026) — is this profile wanted now, or is the stack/role glutted/dying?
5. Rejection modes — specific ways this resume gets killed: ATS, recruiter scan, hiring manager, reference check.
6. Relevance trajectory — where this role/skillset is in 2027-2030 given AI and automation exposure, cited.
7. Positioning critique — is the narrative coherent? Level-appropriate? Or generalist mush / founder-LARP / title-inflated?
8. Red flags — cross-cutting concerns, contradictions, severe issues.
9. What would change the verdict — concrete, observable changes (one specific metric added, one story rewritten, one title corrected) that flip the conclusion.