Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add christyjacob4/claude-tricks --skill investigategit clone --depth 1 https://github.com/christyjacob4/claude-tricksWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/christyjacob4/claude-tricks/investigate)<a href="https://agentmods.dev/skills/christyjacob4/claude-tricks/investigate"><img src="https://agentmods.dev/badge/skills/christyjacob4/claude-tricks/investigate/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/christyjacob4/claude-tricks/investigate"><img src="https://agentmods.dev/badge/skills/christyjacob4/claude-tricks/investigate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00118 | $0.02102 |
| Opus 5 | $0.00059 | $0.01051 |
| Sonnet 5 | $0.00024 | $0.00420 |
| Haiku 4.5 | $0.00012 | $0.00210 |
Grade A, and why
investigate scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 252 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Deep Investigation
Thorough investigation of issues with root cause analysis and recommendations.
OUTPUT: Detailed report with findings, root cause, and recommendations.
Arguments
$ARGUMENTS- Issue description, error message, or area to investigate
Phase 1: Define the Problem (Parallel: Hypotheses + Data Gathering)
Launch two parallel agents immediately. Both can work simultaneously since forming hypotheses from the issue description is independent from gathering raw data.
Agent 1 - Hypotheses Formation:
Analyze the issue description:
$ARGUMENTSForm 3-5 ranked hypotheses about what might be causing this issue. For each hypothesis, state:
- The suspected cause
- Why this is plausible given the symptoms
- What evidence would confirm or refute it
Return a numbered list of hypotheses with supporting reasoning.
Agent 2 - Data Gathering:
Gather all available data about:
$ARGUMENTSCollect in parallel where possible:
- Error messages and stack traces (search logs, output files, CI artifacts)
- Recent changes to affected areas (
git log --oneline -30 -- path/to/affected/)- Related issues or discussions (search issue trackers if available)
- Current state of affected code paths (read the relevant source files)
- Environment and configuration state
Return all raw findings organized by category.
Wait for both agents to complete. Combine their outputs to form the investigation plan for Phase 2.
Phase 2: Evidence Collection (Parallel: 4 Independent Agents)
Launch four parallel agents, one per evidence dimension. These are completely independent investigations that share no dependencies.
Agent 1 - Code Analysis:
Investigate the code paths related to:
$ARGUMENTSUsing the hypotheses: [insert hypotheses from Phase 1]
Analyze:
- All code paths involved in the issue (read source files, trace call chains)
- Recent changes to affected areas (
git log -p --since="2 weeks ago" -- path/to/affected/)- Dependencies and their versions
- Integration points between components
- Any relevant design patterns or architectural decisions that could contribute
For each hypothesis, note which code evidence supports or contradicts it. Return detailed code-level findings with exact file paths and line numbers.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 252 lines · 118 tokens per session scan A 3044bd798ac6
investigate is a skill published in the GitHub repository christyjacob4/claude-tricks (2 stars, last pushed 5mo ago), licensed MIT. It adds 118 tokens to every session and 2,102 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
pero-ponte-sueter
Use when writing code that touches input, I/O, network, parsing, persistence, concurrency, or lifecycle boundaries, to handle the null, empty, malformed, timed-out, partial, and failed paths before the happy path is called finished. NOT for hostile input, injection, or secrets; use ya-te-lavaste-las-manos.
recoge-tu-tiradero
Use after debugging or experimentation, before final evidence or handoff, to remove residue you created that the solution does not need — trace logs, scratch files, .only, temporary flags, abandoned deps. NOT for the handoff gate itself; use vienen-las-visitas. NOT for unreproducible state; use por-si-se-ocupa.
ya-comiste
Use when a failure may come from an unavailable or misconfigured environment — service down, dependency or version missing, unset or expired credential, exhausted port, disk, or memory, unreachable host — before you debug application code. NOT for stale state you created this session; use el-vaporub.
el-vaporub
Use at the very start of a failure investigation, before forming any theory, to rule out your own stale state — unsaved file, unrestarted process, stale cache, wrong branch or worktree, wrong file, outdated build output. Run the ladder once. NOT for an unavailable or misconfigured environment; use ya-comiste.
pero-no-haces-caso
Use after a bug, failure, regression, or rework, to check whether you yourself flagged this exact problem earlier in this session and it was not acted on. Quote your earlier words verbatim or say nothing at all. NOT for late-arriving new scope; use no-se-te-olvide-que.
y-si-lo-encuentro-que
Use before reporting that a repository file, path, symbol, function, route, config key, or dependency is absent or cannot be found. NOT for unverified claims about an external library, API, or platform; use cadena-de-whatsapp.