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 saidsef/mcp-github-pr-issue-analyser --skill error-handlinggit clone --depth 1 https://github.com/saidsef/mcp-github-pr-issue-analyserWrote 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/saidsef/mcp-github-pr-issue-analyser/error-handling)<a href="https://agentmods.dev/skills/saidsef/mcp-github-pr-issue-analyser/error-handling"><img src="https://agentmods.dev/badge/skills/saidsef/mcp-github-pr-issue-analyser/error-handling/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/saidsef/mcp-github-pr-issue-analyser/error-handling"><img src="https://agentmods.dev/badge/skills/saidsef/mcp-github-pr-issue-analyser/error-handling.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00019 | $0.04159 |
| Opus 5.5 | $0.00008 | $0.01664 |
| Sonnet 5.5 | $0.00004 | $0.00832 |
| Haiku 4.5 | $0.00002 | $0.00416 |
Grade A, and why
error-handling 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 21d 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 — 307 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Error Handling
Every tool in this server fails through one small set of typed errors. The code in the message tells you which of the three responses is right: back off, correct the call, or stop and tell the user.
Reading an Error
Failures arrive as a ToolError carrying a bracketed code:
[NOT_FOUND] HTTP 404: PR #4321: Resource not found GitHub said: Not Found
The code is written [<CODE>] HTTP <status>: when a status is known and
[<CODE>] when it is not. Read the code first, since the prose after it
varies with the call and is not stable enough to match on.
The code does not always start the message. A REST call that failed opens with it, as above. An authentication failure, and anything a GraphQL-backed step raised, arrive wrapped once more and open with the tool name instead:
Error calling tool 'github_get_pr_content': [AUTH_FAILED] HTTP 401: PR #1: Authentication failed. Check your GitHub token. GitHub said: Bad credentials
Search the message for the code rather than reading it off the front.
What follows the code has three parts. The call that failed comes first, such
as PR #4321. The server's own reading of the status comes next. Whatever
GitHub said about the refusal comes last, after GitHub said: , and a 422
appends its field-level errors as <field>: <message> after that. The last part
is absent when GitHub sent no body or sent one that is not JSON.
Report the GitHub said text to the user, since it names the cause where the
codes only name the class.
The GraphQL-backed tools label a failed step, so their messages read
[GITHUB_API_ERROR] Failed to <action>: <original error>. The action names the
step that failed, such as fetch user activities or fetch status checks.
Which codes survive that labelling is a question of transport. An error raised
inside the step itself, by the tool or by the GraphQL mapper, keeps its own code
where it is NOT_FOUND or AUTH_FAILED, since the labelling re-raises those
two untouched. github_search_user raises one against its own result for a name
nobody holds:
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.
- 21d ago Changed · +138 lines 677caf99c1ab
- 28d ago Changed 64839cd05e2f
- 1mo ago Changed · +1 lines 8313d33b52e3
- 1mo ago Changed · +37 lines cd3c90313249
- 1mo ago Changed · +27 lines 90234e7309b0
- 1mo ago First seen · 104 lines · 19 tokens per session scan A 3d63b9c86376
error-handling is a skill published in the GitHub repository saidsef/mcp-github-pr-issue-analyser (6 stars, last pushed 2d ago), licensed Apache-2.0. It adds 19 tokens to every session and 4,159 once invoked, about $0.0001 per session on Opus 5.5. 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
agentcore-investigation
Investigate Bedrock AgentCore runtime sessions via CloudWatch Logs Insights — resolve session/trace IDs, query OTEL spans, filter noise, build timelines. Use when debugging AgentCore agent sessions, tracing tool calls, or analyzing latency.
operator-gotchas
MUST READ before wiring a TOP, CHOP, MAT, DAT, light, camera or instancing chain you have not built before: the parameter names TD abbreviates, the defaults that silently do the wrong thing, and the input-order rules, per operator, verified on TD 2025.33230. Pair with describeoptype.
debug-operator
TRIGGER when operator errors appear or user reports broken behavior. Systematic diagnosis workflow.
interacting-with-project-runtime
Interacting with a project's JVM runtime through the Kotlin REPL to execute focused probes against project classes and dependencies. Activate when behavior must be observed by running code; use using-gradle for build inspection or task execution.
watch
Watch live IDX turnover and order books for unusual activity — value surges, auto-reject approach, one-sided depth, walls appearing or withdrawn, quiet stocks waking up, UMA flags — plus delayed print-by-print attribution and end-of-day broker context. Use when the user runs /watch, or asks what is trading right now…
stockbit-status
Check whether the Stockbit session is still valid — main session expiry, whether trading credentials are stored, and the current trading mode. Use when the user asks "am I still logged in to Stockbit", "has my session expired", "is trading on", or when a stockbit MCP tool starts failing with an auth error.