Install
$ agentstack add skill-sigfied-project-to-pdf-resume-skill-project-to-pdf-resume ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Project To PDF Resume
Overview
Create a technical resume from local project folders without inventing impact. Use code evidence for architecture, features, stack, and implementation details; use user confirmation for outcomes, scale, ownership, production results, and business value.
The default output is an evidence-backed resume package: project evidence notes, a claim ledger, outcome questions, polished bullets, source document updates, and a verified PDF when requested.
Progressive Reference Loading
Start with this SKILL.md only. Load bundled references only when the current task reaches the stage that needs that detail; do not preload every reference at the start.
| When the task needs... | Read this reference | | --- | --- | | Stage routing, deliverable overview, or stop conditions | [references/workflow.md](references/workflow.md) | | Intake, project type scanning, code inventory, or first evidence directory | [references/intake-and-scan.md](references/intake-and-scan.md) | | Evidence package artifacts, Markdown evidence docs, JSON metadata/index, claim ledger schema, or traceability rules | [references/evidence-package.md](references/evidence-package.md) | | Architecture synthesis, target-role track selection, material organization, or draft gates | [references/architecture-and-materials.md](references/architecture-and-materials.md) | | Outcome interview questions, claim wording, bullet drafting, or final resume copy review | [references/interview-and-writing.md](references/interview-and-writing.md) | | Source document creation, LaTeX-first PDF production, compilation, rendering, or visual QA | [references/pdf-production.md](references/pdf-production.md) | | Compact Chinese technical resume layout, or a user-provided PDF/TEX layout sample that should guide the final PDF | [references/layout-reference.md](references/layout-reference.md) |
For Simplified Chinese work, load the matching .cn.md counterpart at the same stage.
After reading a reference, apply only the guidance relevant to the current stage, then return to the required flow below.
Required Flow
- Confirm inputs and target.
- Identify the project folders, target role, seniority, resume language, and expected output: notes, resume bullets, source document, PDF, or all of them.
- If the target is unclear, infer a tentative direction from folder names and existing resume files, then state the assumption.
- Preserve private context in task outputs only. Do not copy private project names, company names, file paths, or business details into reusable skill documentation.
- Inventory code before writing.
- First inspect dominant file types and manifests to classify the project type: backend, frontend, data pipeline, mobile app, CLI/tooling, library, infrastructure, or mixed.
- Use that classification to choose the first files to read instead of treating every repository the same.
- Traverse each project folder with
rg --files,find, manifests, README files, config files, and entrypoints. - Read high-signal files: routes, controllers, handlers, services, repositories, jobs, workers, queue/cache code, SQL/query builders, schema or migration code, tests, and deployment files.
- Build a source-of-truth evidence package before writing resume claims.
- Build an evidence package.
- For every project, record role positioning, stack, core modules, request/data flow, technical highlights, implementation details, and evidence files.
- Use Markdown for detailed evidence descriptions and JSON for package metadata/indexes: project index, file evidence index, project evidence card, claim ledger, outcome question backlog, and resume bullet candidate pool when the project has enough substance.
- Label each claim as
Code evidence,Reasonable inference, orNeeds user confirmation. - Treat ownership, scale, production usage, metrics, business value, cost savings, and performance gains as
Needs user confirmationunless they are explicitly documented. - Separate reusable project notes from final resume copy so the same evidence can support multiple role-specific resumes.
- Ask for missing outcomes.
- Ask after the code evidence pass, not before.
- Keep questions short and grounded in observed project behavior.
- Prefer questions about production usage, users, scale, reliability, performance, cost, migration, automation, and the candidate's actual responsibility.
- Ask for exact wording permission when claims may reveal confidential customers, internal systems, or sensitive metrics.
- Draft resume material.
- Match the user's requested language; otherwise use the language of the existing resume.
- Convert evidence into action + method + result bullets.
- Rank bullets by target-role relevance, technical depth, and confirmed outcome strength.
- Keep uncertainty visible in notes; remove unresolved claims from final bullets.
- Produce and verify the PDF when requested.
- Prefer LaTeX for final resume typesetting because it gives stable typography, compact layout control, repeatable builds, and strong PDF output.
- Reuse an existing non-LaTeX source only when it is already the user's canonical resume workflow or the user requests that format.
- Generate or update the source document, compile/render the PDF, inspect page count, render pages to images when possible, and check spacing, overflow, section balance, and font rendering.
- Iterate until the PDF is readable, compact, and truthful.
Rules
- Do not include private project details in reusable skill documentation.
- Do not infer business impact from code alone.
- Prefer project-specific implementation details over generic stack lists.
- Keep uncertainty visible until the user confirms it.
- Preserve existing resume style unless the user asks for a redesign.
- Verify rendered output before claiming a PDF is complete.
- If a claim cannot be verified from code, docs, or user confirmation, keep it out of the final resume.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Sigfied
- Source: Sigfied/project-to-pdf-resume-skill
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.