tam159/next-role/hiring-recon
Claude Code and agent skill by tam159/next-role
- community
Checking that the SKILL.md file is still in the repository.
Pre-interview reconnaissance. Given a processed resume, JD, and optional intake, gather public intel on the company (size, financials, reputation, hiring signals) and the role (market context, salary range bracketed by location), then produce a match analysis as a single markdown report.
Install
npx vskill@latest install tam159/next-role/hiring-reconSKILL.md
Hiring Recon
Pre-interview reconnaissance. After a CV and JD are processed, gather public intelligence on the company and the role and produce a match analysis the candidate can use to prepare.
Inputs
The caller passes exact filesystem paths in the task description:
resume_path— the candidate's processed resume (markdown)jd_path— the processed job description (markdown)intake_path— optional intake notes (may not exist)output_path— where to write the final report
Read every input file in full with read_file(path, limit=1000). Do not skim. Note the candidate's location (from the resume) and the JD's location up front — both matter for salary calibration.
Tools
web_search(query, max_results=5, topic="general"|"news"|"finance")— primary researchweb_extract(urls, content_format="markdown")— pull a single high-value page (careers, About, press release, levels.fyi or local equivalent) when a search snippet is too thin- filesystem tools (
read_file,write_file,edit_file,ls,glob,grep)
Plan 3-5 targeted searches, then optionally one or two web_extract calls. Quality over quantity.
Research axes
- Company snapshot — size, stage (startup / scale-up / public), business model (B2B vs B2C, product vs services/outsourcing), main products, recent news.
- Financial & hiring signals — funding rounds, revenue/profitability hints, layoffs, headcount trend, recent leadership changes. Use
topic="news"ortopic="finance"for these. - Reputation & culture — Glassdoor patterns: report repeated themes in 1-2 star reviews (overwork, attrition, leadership) rather than just the average rating. Tenure signals, controversies.
- Hiring team — if discoverable, who the hiring manager / team lead is and their background. Omit this section entirely if no signal.
- Role market context — typical skills expected in this role family; JD-language red flags using the linguistic taxonomy ("wear many hats" → workload risk; "rockstar/ninja" → culture risk; "competitive salary" → likely below market). Include a Salary range bullet bracketed by location:
Senior <Role>, <region>: ~$X-$Y per <source>. If candidate location and JD location differ, list both. Never quote a generic global number. - Match analysis — given the resume vs. the JD vs. company priorities, strengths to emphasize, gaps to address with adjacent experience. 3-5 bullets each.
Output format
Write a single markdown file to output_path via write_file, with these section headings, in this exact order:
Omit the "Hiring team" section entirely (no header) if you found no signal.
Updates
When the caller's task says "Update the existing report at …" (rather than create one):
read_file(output_path, limit=1000)first — your context is fresh; you have no memory of the prior version.- Identify the surgical change the caller named. The user's explicit request takes priority over the preservation defaults below — if they asked to remove or replace a section, do it; only touch what they named.
- Use
edit_file(path, old_string=..., new_string=...)for targeted insertions or replacements (e.g. adding a subsection under## Hiring team, updating the salary bullet). Usewrite_fileonly when restructuring most of the report. - Preserve every other section, bullet, and source URL the user did not mention. The no-fabrication rule still applies — if a newly-requested fact is not findable, write
Unknown — no public signal. - Reply with the update-mode contract:
Updated research report at: <output_path>.
Rules
- Cite URLs inline next to non-trivial claims. End with a
## Sourceslist. - Truth (absolute): if a fact is not findable, write
Unknown — no public signalrather than fabricating. - By default, keep total under ~600 lines (punchy beats exhaustive) and emit a single file. If the user explicitly asks for a longer report or a different shape, comply.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingbackend/agents/career_agent/skills/hiring-recon/hiring-recon/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/tam159/next-role/hiring-recon)